//===- IR/OpenMPIRBuilder.h - OpenMP encoding builder for LLVM IR - C++ -*-===//
 
//
 
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 
// See https://llvm.org/LICENSE.txt for license information.
 
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 
//
 
//===----------------------------------------------------------------------===//
 
//
 
// This file defines the OpenMPIRBuilder class and helpers used as a convenient
 
// way to create LLVM instructions for OpenMP directives.
 
//
 
//===----------------------------------------------------------------------===//
 
 
 
#ifndef LLVM_FRONTEND_OPENMP_OMPIRBUILDER_H
 
#define LLVM_FRONTEND_OPENMP_OMPIRBUILDER_H
 
 
 
#include "llvm/Analysis/MemorySSAUpdater.h"
 
#include "llvm/Frontend/OpenMP/OMPConstants.h"
 
#include "llvm/IR/DebugLoc.h"
 
#include "llvm/IR/IRBuilder.h"
 
#include "llvm/Support/Allocator.h"
 
#include <forward_list>
 
#include <map>
 
#include <optional>
 
 
 
namespace llvm {
 
class CanonicalLoopInfo;
 
struct TargetRegionEntryInfo;
 
class OffloadEntriesInfoManager;
 
 
 
/// Move the instruction after an InsertPoint to the beginning of another
 
/// BasicBlock.
 
///
 
/// The instructions after \p IP are moved to the beginning of \p New which must
 
/// not have any PHINodes. If \p CreateBranch is true, a branch instruction to
 
/// \p New will be added such that there is no semantic change. Otherwise, the
 
/// \p IP insert block remains degenerate and it is up to the caller to insert a
 
/// terminator.
 
void spliceBB(IRBuilderBase::InsertPoint IP, BasicBlock *New,
 
              bool CreateBranch);
 
 
 
/// Splice a BasicBlock at an IRBuilder's current insertion point. Its new
 
/// insert location will stick to after the instruction before the insertion
 
/// point (instead of moving with the instruction the InsertPoint stores
 
/// internally).
 
void spliceBB(IRBuilder<> &Builder, BasicBlock *New, bool CreateBranch);
 
 
 
/// Split a BasicBlock at an InsertPoint, even if the block is degenerate
 
/// (missing the terminator).
 
///
 
/// llvm::SplitBasicBlock and BasicBlock::splitBasicBlock require a well-formed
 
/// BasicBlock. \p Name is used for the new successor block. If \p CreateBranch
 
/// is true, a branch to the new successor will new created such that
 
/// semantically there is no change; otherwise the block of the insertion point
 
/// remains degenerate and it is the caller's responsibility to insert a
 
/// terminator. Returns the new successor block.
 
BasicBlock *splitBB(IRBuilderBase::InsertPoint IP, bool CreateBranch,
 
                    llvm::Twine Name = {});
 
 
 
/// Split a BasicBlock at \p Builder's insertion point, even if the block is
 
/// degenerate (missing the terminator).  Its new insert location will stick to
 
/// after the instruction before the insertion point (instead of moving with the
 
/// instruction the InsertPoint stores internally).
 
BasicBlock *splitBB(IRBuilderBase &Builder, bool CreateBranch,
 
                    llvm::Twine Name = {});
 
 
 
/// Split a BasicBlock at \p Builder's insertion point, even if the block is
 
/// degenerate (missing the terminator).  Its new insert location will stick to
 
/// after the instruction before the insertion point (instead of moving with the
 
/// instruction the InsertPoint stores internally).
 
BasicBlock *splitBB(IRBuilder<> &Builder, bool CreateBranch, llvm::Twine Name);
 
 
 
/// Like splitBB, but reuses the current block's name for the new name.
 
BasicBlock *splitBBWithSuffix(IRBuilderBase &Builder, bool CreateBranch,
 
                              llvm::Twine Suffix = ".split");
 
 
 
/// Captures attributes that affect generating LLVM-IR using the
 
/// OpenMPIRBuilder and related classes. Note that not all attributes are
 
/// required for all classes or functions. In some use cases the configuration
 
/// is not necessary at all, because because the only functions that are called
 
/// are ones that are not dependent on the configuration.
 
class OpenMPIRBuilderConfig {
 
public:
 
  /// Flag for specifying if the compilation is done for embedded device code
 
  /// or host code.
 
  std::optional<bool> IsEmbedded;
 
 
 
  /// Flag for specifying if the compilation is done for an offloading target,
 
  /// like GPU.
 
  std::optional<bool> IsTargetCodegen;
 
 
 
  /// Flag for specifying weather a requires unified_shared_memory
 
  /// directive is present or not.
 
  std::optional<bool> HasRequiresUnifiedSharedMemory;
 
 
 
  // Flag for specifying if offloading is mandatory.
 
  std::optional<bool> OpenMPOffloadMandatory;
 
 
 
  /// First separator used between the initial two parts of a name.
 
  std::optional<StringRef> FirstSeparator;
 
  /// Separator used between all of the rest consecutive parts of s name
 
  std::optional<StringRef> Separator;
 
 
 
  OpenMPIRBuilderConfig() {}
 
  OpenMPIRBuilderConfig(bool IsEmbedded, bool IsTargetCodegen,
 
                        bool HasRequiresUnifiedSharedMemory,
 
                        bool OpenMPOffloadMandatory)
 
      : IsEmbedded(IsEmbedded), IsTargetCodegen(IsTargetCodegen),
 
        HasRequiresUnifiedSharedMemory(HasRequiresUnifiedSharedMemory),
 
        OpenMPOffloadMandatory(OpenMPOffloadMandatory) {}
 
 
 
  // Getters functions that assert if the required values are not present.
 
  bool isEmbedded() const {
 
    assert(IsEmbedded.has_value() && "IsEmbedded is not set");
 
    return *IsEmbedded;
 
  }
 
 
 
  bool isTargetCodegen() const {
 
    assert(IsTargetCodegen.has_value() && "IsTargetCodegen is not set");
 
    return *IsTargetCodegen;
 
  }
 
 
 
  bool hasRequiresUnifiedSharedMemory() const {
 
    assert(HasRequiresUnifiedSharedMemory.has_value() &&
 
           "HasUnifiedSharedMemory is not set");
 
    return *HasRequiresUnifiedSharedMemory;
 
  }
 
 
 
  bool openMPOffloadMandatory() const {
 
    assert(OpenMPOffloadMandatory.has_value() &&
 
           "OpenMPOffloadMandatory is not set");
 
    return *OpenMPOffloadMandatory;
 
  }
 
  // Returns the FirstSeparator if set, otherwise use the default
 
  // separator depending on isTargetCodegen
 
  StringRef firstSeparator() const {
 
    if (FirstSeparator.has_value())
 
      return *FirstSeparator;
 
    if (isTargetCodegen())
 
      return "_";
 
    return ".";
 
  }
 
 
 
  // Returns the Separator if set, otherwise use the default
 
  // separator depending on isTargetCodegen
 
  StringRef separator() const {
 
    if (Separator.has_value())
 
      return *Separator;
 
    if (isTargetCodegen())
 
      return "$";
 
    return ".";
 
  }
 
 
 
  void setIsEmbedded(bool Value) { IsEmbedded = Value; }
 
  void setIsTargetCodegen(bool Value) { IsTargetCodegen = Value; }
 
  void setHasRequiresUnifiedSharedMemory(bool Value) {
 
    HasRequiresUnifiedSharedMemory = Value;
 
  }
 
  void setFirstSeparator(StringRef FS) { FirstSeparator = FS; }
 
  void setSeparator(StringRef S) { Separator = S; }
 
};
 
 
 
/// An interface to create LLVM-IR for OpenMP directives.
 
///
 
/// Each OpenMP directive has a corresponding public generator method.
 
class OpenMPIRBuilder {
 
public:
 
  /// Create a new OpenMPIRBuilder operating on the given module \p M. This will
 
  /// not have an effect on \p M (see initialize)
 
  OpenMPIRBuilder(Module &M) : M(M), Builder(M.getContext()) {}
 
  ~OpenMPIRBuilder();
 
 
 
  /// Initialize the internal state, this will put structures types and
 
  /// potentially other helpers into the underlying module. Must be called
 
  /// before any other method and only once!
 
  void initialize();
 
 
 
  void setConfig(OpenMPIRBuilderConfig C) { Config = C; }
 
 
 
  /// Finalize the underlying module, e.g., by outlining regions.
 
  /// \param Fn                    The function to be finalized. If not used,
 
  ///                              all functions are finalized.
 
  void finalize(Function *Fn = nullptr);
 
 
 
  /// Add attributes known for \p FnID to \p Fn.
 
  void addAttributes(omp::RuntimeFunction FnID, Function &Fn);
 
 
 
  /// Type used throughout for insertion points.
 
  using InsertPointTy = IRBuilder<>::InsertPoint;
 
 
 
  /// Get the create a name using the platform specific separators.
 
  /// \param Parts parts of the final name that needs separation
 
  /// The created name has a first separator between the first and second part
 
  /// and a second separator between all other parts.
 
  /// E.g. with FirstSeparator "$" and Separator "." and
 
  /// parts: "p1", "p2", "p3", "p4"
 
  /// The resulting name is "p1$p2.p3.p4"
 
  /// The separators are retrieved from the OpenMPIRBuilderConfig.
 
  std::string createPlatformSpecificName(ArrayRef<StringRef> Parts) const;
 
 
 
  /// Callback type for variable finalization (think destructors).
 
  ///
 
  /// \param CodeGenIP is the insertion point at which the finalization code
 
  ///                  should be placed.
 
  ///
 
  /// A finalize callback knows about all objects that need finalization, e.g.
 
  /// destruction, when the scope of the currently generated construct is left
 
  /// at the time, and location, the callback is invoked.
 
  using FinalizeCallbackTy = std::function<void(InsertPointTy CodeGenIP)>;
 
 
 
  struct FinalizationInfo {
 
    /// The finalization callback provided by the last in-flight invocation of
 
    /// createXXXX for the directive of kind DK.
 
    FinalizeCallbackTy FiniCB;
 
 
 
    /// The directive kind of the innermost directive that has an associated
 
    /// region which might require finalization when it is left.
 
    omp::Directive DK;
 
 
 
    /// Flag to indicate if the directive is cancellable.
 
    bool IsCancellable;
 
  };
 
 
 
  /// Push a finalization callback on the finalization stack.
 
  ///
 
  /// NOTE: Temporary solution until Clang CG is gone.
 
  void pushFinalizationCB(const FinalizationInfo &FI) {
 
    FinalizationStack.push_back(FI);
 
  }
 
 
 
  /// Pop the last finalization callback from the finalization stack.
 
  ///
 
  /// NOTE: Temporary solution until Clang CG is gone.
 
  void popFinalizationCB() { FinalizationStack.pop_back(); }
 
 
 
  /// Callback type for body (=inner region) code generation
 
  ///
 
  /// The callback takes code locations as arguments, each describing a
 
  /// location where additional instructions can be inserted.
 
  ///
 
  /// The CodeGenIP may be in the middle of a basic block or point to the end of
 
  /// it. The basic block may have a terminator or be degenerate. The callback
 
  /// function may just insert instructions at that position, but also split the
 
  /// block (without the Before argument of BasicBlock::splitBasicBlock such
 
  /// that the identify of the split predecessor block is preserved) and insert
 
  /// additional control flow, including branches that do not lead back to what
 
  /// follows the CodeGenIP. Note that since the callback is allowed to split
 
  /// the block, callers must assume that InsertPoints to positions in the
 
  /// BasicBlock after CodeGenIP including CodeGenIP itself are invalidated. If
 
  /// such InsertPoints need to be preserved, it can split the block itself
 
  /// before calling the callback.
 
  ///
 
  /// AllocaIP and CodeGenIP must not point to the same position.
 
  ///
 
  /// \param AllocaIP is the insertion point at which new alloca instructions
 
  ///                 should be placed. The BasicBlock it is pointing to must
 
  ///                 not be split.
 
  /// \param CodeGenIP is the insertion point at which the body code should be
 
  ///                  placed.
 
  using BodyGenCallbackTy =
 
      function_ref<void(InsertPointTy AllocaIP, InsertPointTy CodeGenIP)>;
 
 
 
  // This is created primarily for sections construct as llvm::function_ref
 
  // (BodyGenCallbackTy) is not storable (as described in the comments of
 
  // function_ref class - function_ref contains non-ownable reference
 
  // to the callable.
 
  using StorableBodyGenCallbackTy =
 
      std::function<void(InsertPointTy AllocaIP, InsertPointTy CodeGenIP)>;
 
 
 
  /// Callback type for loop body code generation.
 
  ///
 
  /// \param CodeGenIP is the insertion point where the loop's body code must be
 
  ///                  placed. This will be a dedicated BasicBlock with a
 
  ///                  conditional branch from the loop condition check and
 
  ///                  terminated with an unconditional branch to the loop
 
  ///                  latch.
 
  /// \param IndVar    is the induction variable usable at the insertion point.
 
  using LoopBodyGenCallbackTy =
 
      function_ref<void(InsertPointTy CodeGenIP, Value *IndVar)>;
 
 
 
  /// Callback type for variable privatization (think copy & default
 
  /// constructor).
 
  ///
 
  /// \param AllocaIP is the insertion point at which new alloca instructions
 
  ///                 should be placed.
 
  /// \param CodeGenIP is the insertion point at which the privatization code
 
  ///                  should be placed.
 
  /// \param Original The value being copied/created, should not be used in the
 
  ///                 generated IR.
 
  /// \param Inner The equivalent of \p Original that should be used in the
 
  ///              generated IR; this is equal to \p Original if the value is
 
  ///              a pointer and can thus be passed directly, otherwise it is
 
  ///              an equivalent but different value.
 
  /// \param ReplVal The replacement value, thus a copy or new created version
 
  ///                of \p Inner.
 
  ///
 
  /// \returns The new insertion point where code generation continues and
 
  ///          \p ReplVal the replacement value.
 
  using PrivatizeCallbackTy = function_ref<InsertPointTy(
 
      InsertPointTy AllocaIP, InsertPointTy CodeGenIP, Value &Original,
 
      Value &Inner, Value *&ReplVal)>;
 
 
 
  /// Description of a LLVM-IR insertion point (IP) and a debug/source location
 
  /// (filename, line, column, ...).
 
  struct LocationDescription {
 
    LocationDescription(const IRBuilderBase &IRB)
 
        : IP(IRB.saveIP()), DL(IRB.getCurrentDebugLocation()) {}
 
    LocationDescription(const InsertPointTy &IP) : IP(IP) {}
 
    LocationDescription(const InsertPointTy &IP, const DebugLoc &DL)
 
        : IP(IP), DL(DL) {}
 
    InsertPointTy IP;
 
    DebugLoc DL;
 
  };
 
 
 
  /// Emitter methods for OpenMP directives.
 
  ///
 
  ///{
 
 
 
  /// Generator for '#omp barrier'
 
  ///
 
  /// \param Loc The location where the barrier directive was encountered.
 
  /// \param DK The kind of directive that caused the barrier.
 
  /// \param ForceSimpleCall Flag to force a simple (=non-cancellation) barrier.
 
  /// \param CheckCancelFlag Flag to indicate a cancel barrier return value
 
  ///                        should be checked and acted upon.
 
  ///
 
  /// \returns The insertion point after the barrier.
 
  InsertPointTy createBarrier(const LocationDescription &Loc, omp::Directive DK,
 
                              bool ForceSimpleCall = false,
 
                              bool CheckCancelFlag = true);
 
 
 
  /// Generator for '#omp cancel'
 
  ///
 
  /// \param Loc The location where the directive was encountered.
 
  /// \param IfCondition The evaluated 'if' clause expression, if any.
 
  /// \param CanceledDirective The kind of directive that is cancled.
 
  ///
 
  /// \returns The insertion point after the barrier.
 
  InsertPointTy createCancel(const LocationDescription &Loc, Value *IfCondition,
 
                             omp::Directive CanceledDirective);
 
 
 
  /// Generator for '#omp parallel'
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param AllocaIP The insertion points to be used for alloca instructions.
 
  /// \param BodyGenCB Callback that will generate the region code.
 
  /// \param PrivCB Callback to copy a given variable (think copy constructor).
 
  /// \param FiniCB Callback to finalize variable copies.
 
  /// \param IfCondition The evaluated 'if' clause expression, if any.
 
  /// \param NumThreads The evaluated 'num_threads' clause expression, if any.
 
  /// \param ProcBind The value of the 'proc_bind' clause (see ProcBindKind).
 
  /// \param IsCancellable Flag to indicate a cancellable parallel region.
 
  ///
 
  /// \returns The insertion position *after* the parallel.
 
  IRBuilder<>::InsertPoint
 
  createParallel(const LocationDescription &Loc, InsertPointTy AllocaIP,
 
                 BodyGenCallbackTy BodyGenCB, PrivatizeCallbackTy PrivCB,
 
                 FinalizeCallbackTy FiniCB, Value *IfCondition,
 
                 Value *NumThreads, omp::ProcBindKind ProcBind,
 
                 bool IsCancellable);
 
 
 
  /// Generator for the control flow structure of an OpenMP canonical loop.
 
  ///
 
  /// This generator operates on the logical iteration space of the loop, i.e.
 
  /// the caller only has to provide a loop trip count of the loop as defined by
 
  /// base language semantics. The trip count is interpreted as an unsigned
 
  /// integer. The induction variable passed to \p BodyGenCB will be of the same
 
  /// type and run from 0 to \p TripCount - 1. It is up to the callback to
 
  /// convert the logical iteration variable to the loop counter variable in the
 
  /// loop body.
 
  ///
 
  /// \param Loc       The insert and source location description. The insert
 
  ///                  location can be between two instructions or the end of a
 
  ///                  degenerate block (e.g. a BB under construction).
 
  /// \param BodyGenCB Callback that will generate the loop body code.
 
  /// \param TripCount Number of iterations the loop body is executed.
 
  /// \param Name      Base name used to derive BB and instruction names.
 
  ///
 
  /// \returns An object representing the created control flow structure which
 
  ///          can be used for loop-associated directives.
 
  CanonicalLoopInfo *createCanonicalLoop(const LocationDescription &Loc,
 
                                         LoopBodyGenCallbackTy BodyGenCB,
 
                                         Value *TripCount,
 
                                         const Twine &Name = "loop");
 
 
 
  /// Generator for the control flow structure of an OpenMP canonical loop.
 
  ///
 
  /// Instead of a logical iteration space, this allows specifying user-defined
 
  /// loop counter values using increment, upper- and lower bounds. To
 
  /// disambiguate the terminology when counting downwards, instead of lower
 
  /// bounds we use \p Start for the loop counter value in the first body
 
  /// iteration.
 
  ///
 
  /// Consider the following limitations:
 
  ///
 
  ///  * A loop counter space over all integer values of its bit-width cannot be
 
  ///    represented. E.g using uint8_t, its loop trip count of 256 cannot be
 
  ///    stored into an 8 bit integer):
 
  ///
 
  ///      DO I = 0, 255, 1
 
  ///
 
  ///  * Unsigned wrapping is only supported when wrapping only "once"; E.g.
 
  ///    effectively counting downwards:
 
  ///
 
  ///      for (uint8_t i = 100u; i > 0; i += 127u)
 
  ///
 
  ///
 
  /// TODO: May need to add additional parameters to represent:
 
  ///
 
  ///  * Allow representing downcounting with unsigned integers.
 
  ///
 
  ///  * Sign of the step and the comparison operator might disagree:
 
  ///
 
  ///      for (int i = 0; i < 42; i -= 1u)
 
  ///
 
  //
 
  /// \param Loc       The insert and source location description.
 
  /// \param BodyGenCB Callback that will generate the loop body code.
 
  /// \param Start     Value of the loop counter for the first iterations.
 
  /// \param Stop      Loop counter values past this will stop the loop.
 
  /// \param Step      Loop counter increment after each iteration; negative
 
  ///                  means counting down.
 
  /// \param IsSigned  Whether Start, Stop and Step are signed integers.
 
  /// \param InclusiveStop Whether \p Stop itself is a valid value for the loop
 
  ///                      counter.
 
  /// \param ComputeIP Insertion point for instructions computing the trip
 
  ///                  count. Can be used to ensure the trip count is available
 
  ///                  at the outermost loop of a loop nest. If not set,
 
  ///                  defaults to the preheader of the generated loop.
 
  /// \param Name      Base name used to derive BB and instruction names.
 
  ///
 
  /// \returns An object representing the created control flow structure which
 
  ///          can be used for loop-associated directives.
 
  CanonicalLoopInfo *createCanonicalLoop(const LocationDescription &Loc,
 
                                         LoopBodyGenCallbackTy BodyGenCB,
 
                                         Value *Start, Value *Stop, Value *Step,
 
                                         bool IsSigned, bool InclusiveStop,
 
                                         InsertPointTy ComputeIP = {},
 
                                         const Twine &Name = "loop");
 
 
 
  /// Collapse a loop nest into a single loop.
 
  ///
 
  /// Merges loops of a loop nest into a single CanonicalLoopNest representation
 
  /// that has the same number of innermost loop iterations as the origin loop
 
  /// nest. The induction variables of the input loops are derived from the
 
  /// collapsed loop's induction variable. This is intended to be used to
 
  /// implement OpenMP's collapse clause. Before applying a directive,
 
  /// collapseLoops normalizes a loop nest to contain only a single loop and the
 
  /// directive's implementation does not need to handle multiple loops itself.
 
  /// This does not remove the need to handle all loop nest handling by
 
  /// directives, such as the ordered(<n>) clause or the simd schedule-clause
 
  /// modifier of the worksharing-loop directive.
 
  ///
 
  /// Example:
 
  /// \code
 
  ///   for (int i = 0; i < 7; ++i) // Canonical loop "i"
 
  ///     for (int j = 0; j < 9; ++j) // Canonical loop "j"
 
  ///       body(i, j);
 
  /// \endcode
 
  ///
 
  /// After collapsing with Loops={i,j}, the loop is changed to
 
  /// \code
 
  ///   for (int ij = 0; ij < 63; ++ij) {
 
  ///     int i = ij / 9;
 
  ///     int j = ij % 9;
 
  ///     body(i, j);
 
  ///   }
 
  /// \endcode
 
  ///
 
  /// In the current implementation, the following limitations apply:
 
  ///
 
  ///  * All input loops have an induction variable of the same type.
 
  ///
 
  ///  * The collapsed loop will have the same trip count integer type as the
 
  ///    input loops. Therefore it is possible that the collapsed loop cannot
 
  ///    represent all iterations of the input loops. For instance, assuming a
 
  ///    32 bit integer type, and two input loops both iterating 2^16 times, the
 
  ///    theoretical trip count of the collapsed loop would be 2^32 iteration,
 
  ///    which cannot be represented in an 32-bit integer. Behavior is undefined
 
  ///    in this case.
 
  ///
 
  ///  * The trip counts of every input loop must be available at \p ComputeIP.
 
  ///    Non-rectangular loops are not yet supported.
 
  ///
 
  ///  * At each nest level, code between a surrounding loop and its nested loop
 
  ///    is hoisted into the loop body, and such code will be executed more
 
  ///    often than before collapsing (or not at all if any inner loop iteration
 
  ///    has a trip count of 0). This is permitted by the OpenMP specification.
 
  ///
 
  /// \param DL        Debug location for instructions added for collapsing,
 
  ///                  such as instructions to compute/derive the input loop's
 
  ///                  induction variables.
 
  /// \param Loops     Loops in the loop nest to collapse. Loops are specified
 
  ///                  from outermost-to-innermost and every control flow of a
 
  ///                  loop's body must pass through its directly nested loop.
 
  /// \param ComputeIP Where additional instruction that compute the collapsed
 
  ///                  trip count. If not set, defaults to before the generated
 
  ///                  loop.
 
  ///
 
  /// \returns The CanonicalLoopInfo object representing the collapsed loop.
 
  CanonicalLoopInfo *collapseLoops(DebugLoc DL,
 
                                   ArrayRef<CanonicalLoopInfo *> Loops,
 
                                   InsertPointTy ComputeIP);
 
 
 
private:
 
  /// Modifies the canonical loop to be a statically-scheduled workshare loop.
 
  ///
 
  /// This takes a \p LoopInfo representing a canonical loop, such as the one
 
  /// created by \p createCanonicalLoop and emits additional instructions to
 
  /// turn it into a workshare loop. In particular, it calls to an OpenMP
 
  /// runtime function in the preheader to obtain the loop bounds to be used in
 
  /// the current thread, updates the relevant instructions in the canonical
 
  /// loop and calls to an OpenMP runtime finalization function after the loop.
 
  ///
 
  /// \param DL       Debug location for instructions added for the
 
  ///                 workshare-loop construct itself.
 
  /// \param CLI      A descriptor of the canonical loop to workshare.
 
  /// \param AllocaIP An insertion point for Alloca instructions usable in the
 
  ///                 preheader of the loop.
 
  /// \param NeedsBarrier Indicates whether a barrier must be inserted after
 
  ///                     the loop.
 
  ///
 
  /// \returns Point where to insert code after the workshare construct.
 
  InsertPointTy applyStaticWorkshareLoop(DebugLoc DL, CanonicalLoopInfo *CLI,
 
                                         InsertPointTy AllocaIP,
 
                                         bool NeedsBarrier);
 
 
 
  /// Modifies the canonical loop a statically-scheduled workshare loop with a
 
  /// user-specified chunk size.
 
  ///
 
  /// \param DL           Debug location for instructions added for the
 
  ///                     workshare-loop construct itself.
 
  /// \param CLI          A descriptor of the canonical loop to workshare.
 
  /// \param AllocaIP     An insertion point for Alloca instructions usable in
 
  ///                     the preheader of the loop.
 
  /// \param NeedsBarrier Indicates whether a barrier must be inserted after the
 
  ///                     loop.
 
  /// \param ChunkSize    The user-specified chunk size.
 
  ///
 
  /// \returns Point where to insert code after the workshare construct.
 
  InsertPointTy applyStaticChunkedWorkshareLoop(DebugLoc DL,
 
                                                CanonicalLoopInfo *CLI,
 
                                                InsertPointTy AllocaIP,
 
                                                bool NeedsBarrier,
 
                                                Value *ChunkSize);
 
 
 
  /// Modifies the canonical loop to be a dynamically-scheduled workshare loop.
 
  ///
 
  /// This takes a \p LoopInfo representing a canonical loop, such as the one
 
  /// created by \p createCanonicalLoop and emits additional instructions to
 
  /// turn it into a workshare loop. In particular, it calls to an OpenMP
 
  /// runtime function in the preheader to obtain, and then in each iteration
 
  /// to update the loop counter.
 
  ///
 
  /// \param DL       Debug location for instructions added for the
 
  ///                 workshare-loop construct itself.
 
  /// \param CLI      A descriptor of the canonical loop to workshare.
 
  /// \param AllocaIP An insertion point for Alloca instructions usable in the
 
  ///                 preheader of the loop.
 
  /// \param SchedType Type of scheduling to be passed to the init function.
 
  /// \param NeedsBarrier Indicates whether a barrier must be insterted after
 
  ///                     the loop.
 
  /// \param Chunk    The size of loop chunk considered as a unit when
 
  ///                 scheduling. If \p nullptr, defaults to 1.
 
  ///
 
  /// \returns Point where to insert code after the workshare construct.
 
  InsertPointTy applyDynamicWorkshareLoop(DebugLoc DL, CanonicalLoopInfo *CLI,
 
                                          InsertPointTy AllocaIP,
 
                                          omp::OMPScheduleType SchedType,
 
                                          bool NeedsBarrier,
 
                                          Value *Chunk = nullptr);
 
 
 
  /// Create alternative version of the loop to support if clause
 
  ///
 
  /// OpenMP if clause can require to generate second loop. This loop
 
  /// will be executed when if clause condition is not met. createIfVersion
 
  /// adds branch instruction to the copied loop if \p  ifCond is not met.
 
  ///
 
  /// \param Loop       Original loop which should be versioned.
 
  /// \param IfCond     Value which corresponds to if clause condition
 
  /// \param VMap       Value to value map to define relation between
 
  ///                   original and copied loop values and loop blocks.
 
  /// \param NamePrefix Optional name prefix for if.then if.else blocks.
 
  void createIfVersion(CanonicalLoopInfo *Loop, Value *IfCond,
 
                       ValueToValueMapTy &VMap, const Twine &NamePrefix = "");
 
 
 
public:
 
  /// Modifies the canonical loop to be a workshare loop.
 
  ///
 
  /// This takes a \p LoopInfo representing a canonical loop, such as the one
 
  /// created by \p createCanonicalLoop and emits additional instructions to
 
  /// turn it into a workshare loop. In particular, it calls to an OpenMP
 
  /// runtime function in the preheader to obtain the loop bounds to be used in
 
  /// the current thread, updates the relevant instructions in the canonical
 
  /// loop and calls to an OpenMP runtime finalization function after the loop.
 
  ///
 
  /// The concrete transformation is done by applyStaticWorkshareLoop,
 
  /// applyStaticChunkedWorkshareLoop, or applyDynamicWorkshareLoop, depending
 
  /// on the value of \p SchedKind and \p ChunkSize.
 
  ///
 
  /// \param DL       Debug location for instructions added for the
 
  ///                 workshare-loop construct itself.
 
  /// \param CLI      A descriptor of the canonical loop to workshare.
 
  /// \param AllocaIP An insertion point for Alloca instructions usable in the
 
  ///                 preheader of the loop.
 
  /// \param NeedsBarrier Indicates whether a barrier must be insterted after
 
  ///                     the loop.
 
  /// \param SchedKind Scheduling algorithm to use.
 
  /// \param ChunkSize The chunk size for the inner loop.
 
  /// \param HasSimdModifier Whether the simd modifier is present in the
 
  ///                        schedule clause.
 
  /// \param HasMonotonicModifier Whether the monotonic modifier is present in
 
  ///                             the schedule clause.
 
  /// \param HasNonmonotonicModifier Whether the nonmonotonic modifier is
 
  ///                                present in the schedule clause.
 
  /// \param HasOrderedClause Whether the (parameterless) ordered clause is
 
  ///                         present.
 
  ///
 
  /// \returns Point where to insert code after the workshare construct.
 
  InsertPointTy applyWorkshareLoop(
 
      DebugLoc DL, CanonicalLoopInfo *CLI, InsertPointTy AllocaIP,
 
      bool NeedsBarrier,
 
      llvm::omp::ScheduleKind SchedKind = llvm::omp::OMP_SCHEDULE_Default,
 
      Value *ChunkSize = nullptr, bool HasSimdModifier = false,
 
      bool HasMonotonicModifier = false, bool HasNonmonotonicModifier = false,
 
      bool HasOrderedClause = false);
 
 
 
  /// Tile a loop nest.
 
  ///
 
  /// Tiles the loops of \p Loops by the tile sizes in \p TileSizes. Loops in
 
  /// \p/ Loops must be perfectly nested, from outermost to innermost loop
 
  /// (i.e. Loops.front() is the outermost loop). The trip count llvm::Value
 
  /// of every loop and every tile sizes must be usable in the outermost
 
  /// loop's preheader. This implies that the loop nest is rectangular.
 
  ///
 
  /// Example:
 
  /// \code
 
  ///   for (int i = 0; i < 15; ++i) // Canonical loop "i"
 
  ///     for (int j = 0; j < 14; ++j) // Canonical loop "j"
 
  ///         body(i, j);
 
  /// \endcode
 
  ///
 
  /// After tiling with Loops={i,j} and TileSizes={5,7}, the loop is changed to
 
  /// \code
 
  ///   for (int i1 = 0; i1 < 3; ++i1)
 
  ///     for (int j1 = 0; j1 < 2; ++j1)
 
  ///       for (int i2 = 0; i2 < 5; ++i2)
 
  ///         for (int j2 = 0; j2 < 7; ++j2)
 
  ///           body(i1*3+i2, j1*3+j2);
 
  /// \endcode
 
  ///
 
  /// The returned vector are the loops {i1,j1,i2,j2}. The loops i1 and j1 are
 
  /// referred to the floor, and the loops i2 and j2 are the tiles. Tiling also
 
  /// handles non-constant trip counts, non-constant tile sizes and trip counts
 
  /// that are not multiples of the tile size. In the latter case the tile loop
 
  /// of the last floor-loop iteration will have fewer iterations than specified
 
  /// as its tile size.
 
  ///
 
  ///
 
  /// @param DL        Debug location for instructions added by tiling, for
 
  ///                  instance the floor- and tile trip count computation.
 
  /// @param Loops     Loops to tile. The CanonicalLoopInfo objects are
 
  ///                  invalidated by this method, i.e. should not used after
 
  ///                  tiling.
 
  /// @param TileSizes For each loop in \p Loops, the tile size for that
 
  ///                  dimensions.
 
  ///
 
  /// \returns A list of generated loops. Contains twice as many loops as the
 
  ///          input loop nest; the first half are the floor loops and the
 
  ///          second half are the tile loops.
 
  std::vector<CanonicalLoopInfo *>
 
  tileLoops(DebugLoc DL, ArrayRef<CanonicalLoopInfo *> Loops,
 
            ArrayRef<Value *> TileSizes);
 
 
 
  /// Fully unroll a loop.
 
  ///
 
  /// Instead of unrolling the loop immediately (and duplicating its body
 
  /// instructions), it is deferred to LLVM's LoopUnrollPass by adding loop
 
  /// metadata.
 
  ///
 
  /// \param DL   Debug location for instructions added by unrolling.
 
  /// \param Loop The loop to unroll. The loop will be invalidated.
 
  void unrollLoopFull(DebugLoc DL, CanonicalLoopInfo *Loop);
 
 
 
  /// Fully or partially unroll a loop. How the loop is unrolled is determined
 
  /// using LLVM's LoopUnrollPass.
 
  ///
 
  /// \param DL   Debug location for instructions added by unrolling.
 
  /// \param Loop The loop to unroll. The loop will be invalidated.
 
  void unrollLoopHeuristic(DebugLoc DL, CanonicalLoopInfo *Loop);
 
 
 
  /// Partially unroll a loop.
 
  ///
 
  /// The CanonicalLoopInfo of the unrolled loop for use with chained
 
  /// loop-associated directive can be requested using \p UnrolledCLI. Not
 
  /// needing the CanonicalLoopInfo allows more efficient code generation by
 
  /// deferring the actual unrolling to the LoopUnrollPass using loop metadata.
 
  /// A loop-associated directive applied to the unrolled loop needs to know the
 
  /// new trip count which means that if using a heuristically determined unroll
 
  /// factor (\p Factor == 0), that factor must be computed immediately. We are
 
  /// using the same logic as the LoopUnrollPass to derived the unroll factor,
 
  /// but which assumes that some canonicalization has taken place (e.g.
 
  /// Mem2Reg, LICM, GVN, Inlining, etc.). That is, the heuristic will perform
 
  /// better when the unrolled loop's CanonicalLoopInfo is not needed.
 
  ///
 
  /// \param DL          Debug location for instructions added by unrolling.
 
  /// \param Loop        The loop to unroll. The loop will be invalidated.
 
  /// \param Factor      The factor to unroll the loop by. A factor of 0
 
  ///                    indicates that a heuristic should be used to determine
 
  ///                    the unroll-factor.
 
  /// \param UnrolledCLI If non-null, receives the CanonicalLoopInfo of the
 
  ///                    partially unrolled loop. Otherwise, uses loop metadata
 
  ///                    to defer unrolling to the LoopUnrollPass.
 
  void unrollLoopPartial(DebugLoc DL, CanonicalLoopInfo *Loop, int32_t Factor,
 
                         CanonicalLoopInfo **UnrolledCLI);
 
 
 
  /// Add metadata to simd-ize a loop. If IfCond is not nullptr, the loop
 
  /// is cloned. The metadata which prevents vectorization is added to
 
  /// to the cloned loop. The cloned loop is executed when ifCond is evaluated
 
  /// to false.
 
  ///
 
  /// \param Loop        The loop to simd-ize.
 
  /// \param AlignedVars The map which containts pairs of the pointer
 
  ///                    and its corresponding alignment.
 
  /// \param IfCond      The value which corresponds to the if clause
 
  ///                    condition.
 
  /// \param Order       The enum to map order clause.
 
  /// \param Simdlen     The Simdlen length to apply to the simd loop.
 
  /// \param Safelen     The Safelen length to apply to the simd loop.
 
  void applySimd(CanonicalLoopInfo *Loop,
 
                 MapVector<Value *, Value *> AlignedVars, Value *IfCond,
 
                 omp::OrderKind Order, ConstantInt *Simdlen,
 
                 ConstantInt *Safelen);
 
 
 
  /// Generator for '#omp flush'
 
  ///
 
  /// \param Loc The location where the flush directive was encountered
 
  void createFlush(const LocationDescription &Loc);
 
 
 
  /// Generator for '#omp taskwait'
 
  ///
 
  /// \param Loc The location where the taskwait directive was encountered.
 
  void createTaskwait(const LocationDescription &Loc);
 
 
 
  /// Generator for '#omp taskyield'
 
  ///
 
  /// \param Loc The location where the taskyield directive was encountered.
 
  void createTaskyield(const LocationDescription &Loc);
 
 
 
  /// A struct to pack the relevant information for an OpenMP depend clause.
 
  struct DependData {
 
    omp::RTLDependenceKindTy DepKind = omp::RTLDependenceKindTy::DepUnknown;
 
    Type *DepValueType;
 
    Value *DepVal;
 
    explicit DependData() = default;
 
    DependData(omp::RTLDependenceKindTy DepKind, Type *DepValueType,
 
               Value *DepVal)
 
        : DepKind(DepKind), DepValueType(DepValueType), DepVal(DepVal) {}
 
  };
 
 
 
  /// Generator for `#omp task`
 
  ///
 
  /// \param Loc The location where the task construct was encountered.
 
  /// \param AllocaIP The insertion point to be used for alloca instructions.
 
  /// \param BodyGenCB Callback that will generate the region code.
 
  /// \param Tied True if the task is tied, false if the task is untied.
 
  /// \param Final i1 value which is `true` if the task is final, `false` if the
 
  ///              task is not final.
 
  /// \param IfCondition i1 value. If it evaluates to `false`, an undeferred
 
  ///                    task is generated, and the encountering thread must
 
  ///                    suspend the current task region, for which execution
 
  ///                    cannot be resumed until execution of the structured
 
  ///                    block that is associated with the generated task is
 
  ///                    completed.
 
  InsertPointTy createTask(const LocationDescription &Loc,
 
                           InsertPointTy AllocaIP, BodyGenCallbackTy BodyGenCB,
 
                           bool Tied = true, Value *Final = nullptr,
 
                           Value *IfCondition = nullptr,
 
                           SmallVector<DependData> Dependencies = {});
 
 
 
  /// Generator for the taskgroup construct
 
  ///
 
  /// \param Loc The location where the taskgroup construct was encountered.
 
  /// \param AllocaIP The insertion point to be used for alloca instructions.
 
  /// \param BodyGenCB Callback that will generate the region code.
 
  InsertPointTy createTaskgroup(const LocationDescription &Loc,
 
                                InsertPointTy AllocaIP,
 
                                BodyGenCallbackTy BodyGenCB);
 
 
 
  /// Functions used to generate reductions. Such functions take two Values
 
  /// representing LHS and RHS of the reduction, respectively, and a reference
 
  /// to the value that is updated to refer to the reduction result.
 
  using ReductionGenTy =
 
      function_ref<InsertPointTy(InsertPointTy, Value *, Value *, Value *&)>;
 
 
 
  /// Functions used to generate atomic reductions. Such functions take two
 
  /// Values representing pointers to LHS and RHS of the reduction, as well as
 
  /// the element type of these pointers. They are expected to atomically
 
  /// update the LHS to the reduced value.
 
  using AtomicReductionGenTy =
 
      function_ref<InsertPointTy(InsertPointTy, Type *, Value *, Value *)>;
 
 
 
  /// Information about an OpenMP reduction.
 
  struct ReductionInfo {
 
    ReductionInfo(Type *ElementType, Value *Variable, Value *PrivateVariable,
 
                  ReductionGenTy ReductionGen,
 
                  AtomicReductionGenTy AtomicReductionGen)
 
        : ElementType(ElementType), Variable(Variable),
 
          PrivateVariable(PrivateVariable), ReductionGen(ReductionGen),
 
          AtomicReductionGen(AtomicReductionGen) {
 
      assert(cast<PointerType>(Variable->getType())
 
          ->isOpaqueOrPointeeTypeMatches(ElementType) && "Invalid elem type");
 
    }
 
 
 
    /// Reduction element type, must match pointee type of variable.
 
    Type *ElementType;
 
 
 
    /// Reduction variable of pointer type.
 
    Value *Variable;
 
 
 
    /// Thread-private partial reduction variable.
 
    Value *PrivateVariable;
 
 
 
    /// Callback for generating the reduction body. The IR produced by this will
 
    /// be used to combine two values in a thread-safe context, e.g., under
 
    /// lock or within the same thread, and therefore need not be atomic.
 
    ReductionGenTy ReductionGen;
 
 
 
    /// Callback for generating the atomic reduction body, may be null. The IR
 
    /// produced by this will be used to atomically combine two values during
 
    /// reduction. If null, the implementation will use the non-atomic version
 
    /// along with the appropriate synchronization mechanisms.
 
    AtomicReductionGenTy AtomicReductionGen;
 
  };
 
 
 
  // TODO: provide atomic and non-atomic reduction generators for reduction
 
  // operators defined by the OpenMP specification.
 
 
 
  /// Generator for '#omp reduction'.
 
  ///
 
  /// Emits the IR instructing the runtime to perform the specific kind of
 
  /// reductions. Expects reduction variables to have been privatized and
 
  /// initialized to reduction-neutral values separately. Emits the calls to
 
  /// runtime functions as well as the reduction function and the basic blocks
 
  /// performing the reduction atomically and non-atomically.
 
  ///
 
  /// The code emitted for the following:
 
  ///
 
  /// \code
 
  ///   type var_1;
 
  ///   type var_2;
 
  ///   #pragma omp <directive> reduction(reduction-op:var_1,var_2)
 
  ///   /* body */;
 
  /// \endcode
 
  ///
 
  /// corresponds to the following sketch.
 
  ///
 
  /// \code
 
  /// void _outlined_par() {
 
  ///   // N is the number of different reductions.
 
  ///   void *red_array[] = {privatized_var_1, privatized_var_2, ...};
 
  ///   switch(__kmpc_reduce(..., N, /*size of data in red array*/, red_array,
 
  ///                        _omp_reduction_func,
 
  ///                        _gomp_critical_user.reduction.var)) {
 
  ///   case 1: {
 
  ///     var_1 = var_1 <reduction-op> privatized_var_1;
 
  ///     var_2 = var_2 <reduction-op> privatized_var_2;
 
  ///     // ...
 
  ///    __kmpc_end_reduce(...);
 
  ///     break;
 
  ///   }
 
  ///   case 2: {
 
  ///     _Atomic<ReductionOp>(var_1, privatized_var_1);
 
  ///     _Atomic<ReductionOp>(var_2, privatized_var_2);
 
  ///     // ...
 
  ///     break;
 
  ///   }
 
  ///   default: break;
 
  ///   }
 
  /// }
 
  ///
 
  /// void _omp_reduction_func(void **lhs, void **rhs) {
 
  ///   *(type *)lhs[0] = *(type *)lhs[0] <reduction-op> *(type *)rhs[0];
 
  ///   *(type *)lhs[1] = *(type *)lhs[1] <reduction-op> *(type *)rhs[1];
 
  ///   // ...
 
  /// }
 
  /// \endcode
 
  ///
 
  /// \param Loc                The location where the reduction was
 
  ///                           encountered. Must be within the associate
 
  ///                           directive and after the last local access to the
 
  ///                           reduction variables.
 
  /// \param AllocaIP           An insertion point suitable for allocas usable
 
  ///                           in reductions.
 
  /// \param ReductionInfos     A list of info on each reduction variable.
 
  /// \param IsNoWait           A flag set if the reduction is marked as nowait.
 
  InsertPointTy createReductions(const LocationDescription &Loc,
 
                                 InsertPointTy AllocaIP,
 
                                 ArrayRef<ReductionInfo> ReductionInfos,
 
                                 bool IsNoWait = false);
 
 
 
  ///}
 
 
 
  /// Return the insertion point used by the underlying IRBuilder.
 
  InsertPointTy getInsertionPoint() { return Builder.saveIP(); }
 
 
 
  /// Update the internal location to \p Loc.
 
  bool updateToLocation(const LocationDescription &Loc) {
 
    Builder.restoreIP(Loc.IP);
 
    Builder.SetCurrentDebugLocation(Loc.DL);
 
    return Loc.IP.getBlock() != nullptr;
 
  }
 
 
 
  /// Return the function declaration for the runtime function with \p FnID.
 
  FunctionCallee getOrCreateRuntimeFunction(Module &M,
 
                                            omp::RuntimeFunction FnID);
 
 
 
  Function *getOrCreateRuntimeFunctionPtr(omp::RuntimeFunction FnID);
 
 
 
  /// Return the (LLVM-IR) string describing the source location \p LocStr.
 
  Constant *getOrCreateSrcLocStr(StringRef LocStr, uint32_t &SrcLocStrSize);
 
 
 
  /// Return the (LLVM-IR) string describing the default source location.
 
  Constant *getOrCreateDefaultSrcLocStr(uint32_t &SrcLocStrSize);
 
 
 
  /// Return the (LLVM-IR) string describing the source location identified by
 
  /// the arguments.
 
  Constant *getOrCreateSrcLocStr(StringRef FunctionName, StringRef FileName,
 
                                 unsigned Line, unsigned Column,
 
                                 uint32_t &SrcLocStrSize);
 
 
 
  /// Return the (LLVM-IR) string describing the DebugLoc \p DL. Use \p F as
 
  /// fallback if \p DL does not specify the function name.
 
  Constant *getOrCreateSrcLocStr(DebugLoc DL, uint32_t &SrcLocStrSize,
 
                                 Function *F = nullptr);
 
 
 
  /// Return the (LLVM-IR) string describing the source location \p Loc.
 
  Constant *getOrCreateSrcLocStr(const LocationDescription &Loc,
 
                                 uint32_t &SrcLocStrSize);
 
 
 
  /// Return an ident_t* encoding the source location \p SrcLocStr and \p Flags.
 
  /// TODO: Create a enum class for the Reserve2Flags
 
  Constant *getOrCreateIdent(Constant *SrcLocStr, uint32_t SrcLocStrSize,
 
                             omp::IdentFlag Flags = omp::IdentFlag(0),
 
                             unsigned Reserve2Flags = 0);
 
 
 
  /// Create a hidden global flag \p Name in the module with initial value \p
 
  /// Value.
 
  GlobalValue *createGlobalFlag(unsigned Value, StringRef Name);
 
 
 
  /// Create an offloading section struct used to register this global at
 
  /// runtime.
 
  ///
 
  /// Type struct __tgt_offload_entry{
 
  ///   void    *addr;      // Pointer to the offload entry info.
 
  ///                       // (function or global)
 
  ///   char    *name;      // Name of the function or global.
 
  ///   size_t  size;       // Size of the entry info (0 if it a function).
 
  ///   int32_t flags;
 
  ///   int32_t reserved;
 
  /// };
 
  ///
 
  /// \param Addr The pointer to the global being registered.
 
  /// \param Name The symbol name associated with the global.
 
  /// \param Size The size in bytes of the global (0 for functions).
 
  /// \param Flags Flags associated with the entry.
 
  /// \param SectionName The section this entry will be placed at.
 
  void emitOffloadingEntry(Constant *Addr, StringRef Name, uint64_t Size,
 
                           int32_t Flags,
 
                           StringRef SectionName = "omp_offloading_entries");
 
 
 
  /// Generate control flow and cleanup for cancellation.
 
  ///
 
  /// \param CancelFlag Flag indicating if the cancellation is performed.
 
  /// \param CanceledDirective The kind of directive that is cancled.
 
  /// \param ExitCB Extra code to be generated in the exit block.
 
  void emitCancelationCheckImpl(Value *CancelFlag,
 
                                omp::Directive CanceledDirective,
 
                                FinalizeCallbackTy ExitCB = {});
 
 
 
  /// Generate a target region entry call.
 
  ///
 
  /// \param Loc The location at which the request originated and is fulfilled.
 
  /// \param Return Return value of the created function returned by reference.
 
  /// \param DeviceID Identifier for the device via the 'device' clause.
 
  /// \param NumTeams Numer of teams for the region via the 'num_teams' clause
 
  ///                 or 0 if unspecified and -1 if there is no 'teams' clause.
 
  /// \param NumThreads Number of threads via the 'thread_limit' clause.
 
  /// \param HostPtr Pointer to the host-side pointer of the target kernel.
 
  /// \param KernelArgs Array of arguments to the kernel.
 
  InsertPointTy emitTargetKernel(const LocationDescription &Loc, Value *&Return,
 
                                 Value *Ident, Value *DeviceID, Value *NumTeams,
 
                                 Value *NumThreads, Value *HostPtr,
 
                                 ArrayRef<Value *> KernelArgs);
 
 
 
  /// Generate a barrier runtime call.
 
  ///
 
  /// \param Loc The location at which the request originated and is fulfilled.
 
  /// \param DK The directive which caused the barrier
 
  /// \param ForceSimpleCall Flag to force a simple (=non-cancellation) barrier.
 
  /// \param CheckCancelFlag Flag to indicate a cancel barrier return value
 
  ///                        should be checked and acted upon.
 
  ///
 
  /// \returns The insertion point after the barrier.
 
  InsertPointTy emitBarrierImpl(const LocationDescription &Loc,
 
                                omp::Directive DK, bool ForceSimpleCall,
 
                                bool CheckCancelFlag);
 
 
 
  /// Generate a flush runtime call.
 
  ///
 
  /// \param Loc The location at which the request originated and is fulfilled.
 
  void emitFlush(const LocationDescription &Loc);
 
 
 
  /// The finalization stack made up of finalize callbacks currently in-flight,
 
  /// wrapped into FinalizationInfo objects that reference also the finalization
 
  /// target block and the kind of cancellable directive.
 
  SmallVector<FinalizationInfo, 8> FinalizationStack;
 
 
 
  /// Return true if the last entry in the finalization stack is of kind \p DK
 
  /// and cancellable.
 
  bool isLastFinalizationInfoCancellable(omp::Directive DK) {
 
    return !FinalizationStack.empty() &&
 
           FinalizationStack.back().IsCancellable &&
 
           FinalizationStack.back().DK == DK;
 
  }
 
 
 
  /// Generate a taskwait runtime call.
 
  ///
 
  /// \param Loc The location at which the request originated and is fulfilled.
 
  void emitTaskwaitImpl(const LocationDescription &Loc);
 
 
 
  /// Generate a taskyield runtime call.
 
  ///
 
  /// \param Loc The location at which the request originated and is fulfilled.
 
  void emitTaskyieldImpl(const LocationDescription &Loc);
 
 
 
  /// Return the current thread ID.
 
  ///
 
  /// \param Ident The ident (ident_t*) describing the query origin.
 
  Value *getOrCreateThreadID(Value *Ident);
 
 
 
  /// The OpenMPIRBuilder Configuration
 
  OpenMPIRBuilderConfig Config;
 
 
 
  /// The underlying LLVM-IR module
 
  Module &M;
 
 
 
  /// The LLVM-IR Builder used to create IR.
 
  IRBuilder<> Builder;
 
 
 
  /// Map to remember source location strings
 
  StringMap<Constant *> SrcLocStrMap;
 
 
 
  /// Map to remember existing ident_t*.
 
  DenseMap<std::pair<Constant *, uint64_t>, Constant *> IdentMap;
 
 
 
  /// Helper that contains information about regions we need to outline
 
  /// during finalization.
 
  struct OutlineInfo {
 
    using PostOutlineCBTy = std::function<void(Function &)>;
 
    PostOutlineCBTy PostOutlineCB;
 
    BasicBlock *EntryBB, *ExitBB, *OuterAllocaBB;
 
    SmallVector<Value *, 2> ExcludeArgsFromAggregate;
 
 
 
    /// Collect all blocks in between EntryBB and ExitBB in both the given
 
    /// vector and set.
 
    void collectBlocks(SmallPtrSetImpl<BasicBlock *> &BlockSet,
 
                       SmallVectorImpl<BasicBlock *> &BlockVector);
 
 
 
    /// Return the function that contains the region to be outlined.
 
    Function *getFunction() const { return EntryBB->getParent(); }
 
  };
 
 
 
  /// Collection of regions that need to be outlined during finalization.
 
  SmallVector<OutlineInfo, 16> OutlineInfos;
 
 
 
  /// Collection of owned canonical loop objects that eventually need to be
 
  /// free'd.
 
  std::forward_list<CanonicalLoopInfo> LoopInfos;
 
 
 
  /// Add a new region that will be outlined later.
 
  void addOutlineInfo(OutlineInfo &&OI) { OutlineInfos.emplace_back(OI); }
 
 
 
  /// An ordered map of auto-generated variables to their unique names.
 
  /// It stores variables with the following names: 1) ".gomp_critical_user_" +
 
  /// <critical_section_name> + ".var" for "omp critical" directives; 2)
 
  /// <mangled_name_for_global_var> + ".cache." for cache for threadprivate
 
  /// variables.
 
  StringMap<Constant*, BumpPtrAllocator> InternalVars;
 
 
 
  /// Create the global variable holding the offload mappings information.
 
  GlobalVariable *createOffloadMaptypes(SmallVectorImpl<uint64_t> &Mappings,
 
                                        std::string VarName);
 
 
 
  /// Create the global variable holding the offload names information.
 
  GlobalVariable *
 
  createOffloadMapnames(SmallVectorImpl<llvm::Constant *> &Names,
 
                        std::string VarName);
 
 
 
  struct MapperAllocas {
 
    AllocaInst *ArgsBase = nullptr;
 
    AllocaInst *Args = nullptr;
 
    AllocaInst *ArgSizes = nullptr;
 
  };
 
 
 
  /// Create the allocas instruction used in call to mapper functions.
 
  void createMapperAllocas(const LocationDescription &Loc,
 
                           InsertPointTy AllocaIP, unsigned NumOperands,
 
                           struct MapperAllocas &MapperAllocas);
 
 
 
  /// Create the call for the target mapper function.
 
  /// \param Loc The source location description.
 
  /// \param MapperFunc Function to be called.
 
  /// \param SrcLocInfo Source location information global.
 
  /// \param MaptypesArg The argument types.
 
  /// \param MapnamesArg The argument names.
 
  /// \param MapperAllocas The AllocaInst used for the call.
 
  /// \param DeviceID Device ID for the call.
 
  /// \param NumOperands Number of operands in the call.
 
  void emitMapperCall(const LocationDescription &Loc, Function *MapperFunc,
 
                      Value *SrcLocInfo, Value *MaptypesArg, Value *MapnamesArg,
 
                      struct MapperAllocas &MapperAllocas, int64_t DeviceID,
 
                      unsigned NumOperands);
 
 
 
  /// Container for the arguments used to pass data to the runtime library.
 
  struct TargetDataRTArgs {
 
    explicit TargetDataRTArgs() {}
 
    /// The array of base pointer passed to the runtime library.
 
    Value *BasePointersArray = nullptr;
 
    /// The array of section pointers passed to the runtime library.
 
    Value *PointersArray = nullptr;
 
    /// The array of sizes passed to the runtime library.
 
    Value *SizesArray = nullptr;
 
    /// The array of map types passed to the runtime library for the beginning
 
    /// of the region or for the entire region if there are no separate map
 
    /// types for the region end.
 
    Value *MapTypesArray = nullptr;
 
    /// The array of map types passed to the runtime library for the end of the
 
    /// region, or nullptr if there are no separate map types for the region
 
    /// end.
 
    Value *MapTypesArrayEnd = nullptr;
 
    /// The array of user-defined mappers passed to the runtime library.
 
    Value *MappersArray = nullptr;
 
    /// The array of original declaration names of mapped pointers sent to the
 
    /// runtime library for debugging
 
    Value *MapNamesArray = nullptr;
 
  };
 
 
 
  /// Struct that keeps the information that should be kept throughout
 
  /// a 'target data' region.
 
  class TargetDataInfo {
 
    /// Set to true if device pointer information have to be obtained.
 
    bool RequiresDevicePointerInfo = false;
 
    /// Set to true if Clang emits separate runtime calls for the beginning and
 
    /// end of the region.  These calls might have separate map type arrays.
 
    bool SeparateBeginEndCalls = false;
 
 
 
  public:
 
    TargetDataRTArgs RTArgs;
 
 
 
    /// Indicate whether any user-defined mapper exists.
 
    bool HasMapper = false;
 
    /// The total number of pointers passed to the runtime library.
 
    unsigned NumberOfPtrs = 0u;
 
 
 
    explicit TargetDataInfo() {}
 
    explicit TargetDataInfo(bool RequiresDevicePointerInfo,
 
                            bool SeparateBeginEndCalls)
 
        : RequiresDevicePointerInfo(RequiresDevicePointerInfo),
 
          SeparateBeginEndCalls(SeparateBeginEndCalls) {}
 
    /// Clear information about the data arrays.
 
    void clearArrayInfo() {
 
      RTArgs = TargetDataRTArgs();
 
      HasMapper = false;
 
      NumberOfPtrs = 0u;
 
    }
 
    /// Return true if the current target data information has valid arrays.
 
    bool isValid() {
 
      return RTArgs.BasePointersArray && RTArgs.PointersArray &&
 
             RTArgs.SizesArray && RTArgs.MapTypesArray &&
 
             (!HasMapper || RTArgs.MappersArray) && NumberOfPtrs;
 
    }
 
    bool requiresDevicePointerInfo() { return RequiresDevicePointerInfo; }
 
    bool separateBeginEndCalls() { return SeparateBeginEndCalls; }
 
  };
 
 
 
  /// Emit the arguments to be passed to the runtime library based on the
 
  /// arrays of base pointers, pointers, sizes, map types, and mappers.  If
 
  /// ForEndCall, emit map types to be passed for the end of the region instead
 
  /// of the beginning.
 
  void emitOffloadingArraysArgument(IRBuilderBase &Builder,
 
                                    OpenMPIRBuilder::TargetDataRTArgs &RTArgs,
 
                                    OpenMPIRBuilder::TargetDataInfo &Info,
 
                                    bool EmitDebug = false,
 
                                    bool ForEndCall = false);
 
 
 
  /// Creates offloading entry for the provided entry ID \a ID, address \a
 
  /// Addr, size \a Size, and flags \a Flags.
 
  void createOffloadEntry(Constant *ID, Constant *Addr, uint64_t Size,
 
                          int32_t Flags, GlobalValue::LinkageTypes);
 
 
 
  /// The kind of errors that can occur when emitting the offload entries and
 
  /// metadata.
 
  enum EmitMetadataErrorKind {
 
    EMIT_MD_TARGET_REGION_ERROR,
 
    EMIT_MD_DECLARE_TARGET_ERROR,
 
    EMIT_MD_GLOBAL_VAR_LINK_ERROR
 
  };
 
 
 
  /// Callback function type
 
  using EmitMetadataErrorReportFunctionTy =
 
      std::function<void(EmitMetadataErrorKind, TargetRegionEntryInfo)>;
 
 
 
  // Emit the offloading entries and metadata so that the device codegen side
 
  // can easily figure out what to emit. The produced metadata looks like
 
  // this:
 
  //
 
  // !omp_offload.info = !{!1, ...}
 
  //
 
  // We only generate metadata for function that contain target regions.
 
  void createOffloadEntriesAndInfoMetadata(
 
      OffloadEntriesInfoManager &OffloadEntriesInfoManager,
 
      EmitMetadataErrorReportFunctionTy &ErrorReportFunction);
 
 
 
public:
 
  /// Generator for __kmpc_copyprivate
 
  ///
 
  /// \param Loc The source location description.
 
  /// \param BufSize Number of elements in the buffer.
 
  /// \param CpyBuf List of pointers to data to be copied.
 
  /// \param CpyFn function to call for copying data.
 
  /// \param DidIt flag variable; 1 for 'single' thread, 0 otherwise.
 
  ///
 
  /// \return The insertion position *after* the CopyPrivate call.
 
 
 
  InsertPointTy createCopyPrivate(const LocationDescription &Loc,
 
                                  llvm::Value *BufSize, llvm::Value *CpyBuf,
 
                                  llvm::Value *CpyFn, llvm::Value *DidIt);
 
 
 
  /// Generator for '#omp single'
 
  ///
 
  /// \param Loc The source location description.
 
  /// \param BodyGenCB Callback that will generate the region code.
 
  /// \param FiniCB Callback to finalize variable copies.
 
  /// \param IsNowait If false, a barrier is emitted.
 
  /// \param DidIt Local variable used as a flag to indicate 'single' thread
 
  ///
 
  /// \returns The insertion position *after* the single call.
 
  InsertPointTy createSingle(const LocationDescription &Loc,
 
                             BodyGenCallbackTy BodyGenCB,
 
                             FinalizeCallbackTy FiniCB, bool IsNowait,
 
                             llvm::Value *DidIt);
 
 
 
  /// Generator for '#omp master'
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param BodyGenCB Callback that will generate the region code.
 
  /// \param FiniCB Callback to finalize variable copies.
 
  ///
 
  /// \returns The insertion position *after* the master.
 
  InsertPointTy createMaster(const LocationDescription &Loc,
 
                             BodyGenCallbackTy BodyGenCB,
 
                             FinalizeCallbackTy FiniCB);
 
 
 
  /// Generator for '#omp masked'
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param BodyGenCB Callback that will generate the region code.
 
  /// \param FiniCB Callback to finialize variable copies.
 
  ///
 
  /// \returns The insertion position *after* the masked.
 
  InsertPointTy createMasked(const LocationDescription &Loc,
 
                             BodyGenCallbackTy BodyGenCB,
 
                             FinalizeCallbackTy FiniCB, Value *Filter);
 
 
 
  /// Generator for '#omp critical'
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param BodyGenCB Callback that will generate the region body code.
 
  /// \param FiniCB Callback to finalize variable copies.
 
  /// \param CriticalName name of the lock used by the critical directive
 
  /// \param HintInst Hint Instruction for hint clause associated with critical
 
  ///
 
  /// \returns The insertion position *after* the critical.
 
  InsertPointTy createCritical(const LocationDescription &Loc,
 
                               BodyGenCallbackTy BodyGenCB,
 
                               FinalizeCallbackTy FiniCB,
 
                               StringRef CriticalName, Value *HintInst);
 
 
 
  /// Generator for '#omp ordered depend (source | sink)'
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param AllocaIP The insertion point to be used for alloca instructions.
 
  /// \param NumLoops The number of loops in depend clause.
 
  /// \param StoreValues The value will be stored in vector address.
 
  /// \param Name The name of alloca instruction.
 
  /// \param IsDependSource If true, depend source; otherwise, depend sink.
 
  ///
 
  /// \return The insertion position *after* the ordered.
 
  InsertPointTy createOrderedDepend(const LocationDescription &Loc,
 
                                    InsertPointTy AllocaIP, unsigned NumLoops,
 
                                    ArrayRef<llvm::Value *> StoreValues,
 
                                    const Twine &Name, bool IsDependSource);
 
 
 
  /// Generator for '#omp ordered [threads | simd]'
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param BodyGenCB Callback that will generate the region code.
 
  /// \param FiniCB Callback to finalize variable copies.
 
  /// \param IsThreads If true, with threads clause or without clause;
 
  /// otherwise, with simd clause;
 
  ///
 
  /// \returns The insertion position *after* the ordered.
 
  InsertPointTy createOrderedThreadsSimd(const LocationDescription &Loc,
 
                                         BodyGenCallbackTy BodyGenCB,
 
                                         FinalizeCallbackTy FiniCB,
 
                                         bool IsThreads);
 
 
 
  /// Generator for '#omp sections'
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param AllocaIP The insertion points to be used for alloca instructions.
 
  /// \param SectionCBs Callbacks that will generate body of each section.
 
  /// \param PrivCB Callback to copy a given variable (think copy constructor).
 
  /// \param FiniCB Callback to finalize variable copies.
 
  /// \param IsCancellable Flag to indicate a cancellable parallel region.
 
  /// \param IsNowait If true, barrier - to ensure all sections are executed
 
  /// before moving forward will not be generated.
 
  /// \returns The insertion position *after* the sections.
 
  InsertPointTy createSections(const LocationDescription &Loc,
 
                               InsertPointTy AllocaIP,
 
                               ArrayRef<StorableBodyGenCallbackTy> SectionCBs,
 
                               PrivatizeCallbackTy PrivCB,
 
                               FinalizeCallbackTy FiniCB, bool IsCancellable,
 
                               bool IsNowait);
 
 
 
  /// Generator for '#omp section'
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param BodyGenCB Callback that will generate the region body code.
 
  /// \param FiniCB Callback to finalize variable copies.
 
  /// \returns The insertion position *after* the section.
 
  InsertPointTy createSection(const LocationDescription &Loc,
 
                              BodyGenCallbackTy BodyGenCB,
 
                              FinalizeCallbackTy FiniCB);
 
 
 
  /// Generate conditional branch and relevant BasicBlocks through which private
 
  /// threads copy the 'copyin' variables from Master copy to threadprivate
 
  /// copies.
 
  ///
 
  /// \param IP insertion block for copyin conditional
 
  /// \param MasterVarPtr a pointer to the master variable
 
  /// \param PrivateVarPtr a pointer to the threadprivate variable
 
  /// \param IntPtrTy Pointer size type
 
  /// \param BranchtoEnd Create a branch between the copyin.not.master blocks
 
  //                             and copy.in.end block
 
  ///
 
  /// \returns The insertion point where copying operation to be emitted.
 
  InsertPointTy createCopyinClauseBlocks(InsertPointTy IP, Value *MasterAddr,
 
                                         Value *PrivateAddr,
 
                                         llvm::IntegerType *IntPtrTy,
 
                                         bool BranchtoEnd = true);
 
 
 
  /// Create a runtime call for kmpc_Alloc
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param Size Size of allocated memory space
 
  /// \param Allocator Allocator information instruction
 
  /// \param Name Name of call Instruction for OMP_alloc
 
  ///
 
  /// \returns CallInst to the OMP_Alloc call
 
  CallInst *createOMPAlloc(const LocationDescription &Loc, Value *Size,
 
                           Value *Allocator, std::string Name = "");
 
 
 
  /// Create a runtime call for kmpc_free
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param Addr Address of memory space to be freed
 
  /// \param Allocator Allocator information instruction
 
  /// \param Name Name of call Instruction for OMP_Free
 
  ///
 
  /// \returns CallInst to the OMP_Free call
 
  CallInst *createOMPFree(const LocationDescription &Loc, Value *Addr,
 
                          Value *Allocator, std::string Name = "");
 
 
 
  /// Create a runtime call for kmpc_threadprivate_cached
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param Pointer pointer to data to be cached
 
  /// \param Size size of data to be cached
 
  /// \param Name Name of call Instruction for callinst
 
  ///
 
  /// \returns CallInst to the thread private cache call.
 
  CallInst *createCachedThreadPrivate(const LocationDescription &Loc,
 
                                      llvm::Value *Pointer,
 
                                      llvm::ConstantInt *Size,
 
                                      const llvm::Twine &Name = Twine(""));
 
 
 
  /// Create a runtime call for __tgt_interop_init
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param InteropVar variable to be allocated
 
  /// \param InteropType type of interop operation
 
  /// \param Device devide to which offloading will occur
 
  /// \param NumDependences  number of dependence variables
 
  /// \param DependenceAddress pointer to dependence variables
 
  /// \param HaveNowaitClause does nowait clause exist
 
  ///
 
  /// \returns CallInst to the __tgt_interop_init call
 
  CallInst *createOMPInteropInit(const LocationDescription &Loc,
 
                                 Value *InteropVar,
 
                                 omp::OMPInteropType InteropType, Value *Device,
 
                                 Value *NumDependences,
 
                                 Value *DependenceAddress,
 
                                 bool HaveNowaitClause);
 
 
 
  /// Create a runtime call for __tgt_interop_destroy
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param InteropVar variable to be allocated
 
  /// \param Device devide to which offloading will occur
 
  /// \param NumDependences  number of dependence variables
 
  /// \param DependenceAddress pointer to dependence variables
 
  /// \param HaveNowaitClause does nowait clause exist
 
  ///
 
  /// \returns CallInst to the __tgt_interop_destroy call
 
  CallInst *createOMPInteropDestroy(const LocationDescription &Loc,
 
                                    Value *InteropVar, Value *Device,
 
                                    Value *NumDependences,
 
                                    Value *DependenceAddress,
 
                                    bool HaveNowaitClause);
 
 
 
  /// Create a runtime call for __tgt_interop_use
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param InteropVar variable to be allocated
 
  /// \param Device devide to which offloading will occur
 
  /// \param NumDependences  number of dependence variables
 
  /// \param DependenceAddress pointer to dependence variables
 
  /// \param HaveNowaitClause does nowait clause exist
 
  ///
 
  /// \returns CallInst to the __tgt_interop_use call
 
  CallInst *createOMPInteropUse(const LocationDescription &Loc,
 
                                Value *InteropVar, Value *Device,
 
                                Value *NumDependences, Value *DependenceAddress,
 
                                bool HaveNowaitClause);
 
 
 
  /// The `omp target` interface
 
  ///
 
  /// For more information about the usage of this interface,
 
  /// \see openmp/libomptarget/deviceRTLs/common/include/target.h
 
  ///
 
  ///{
 
 
 
  /// Create a runtime call for kmpc_target_init
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param IsSPMD Flag to indicate if the kernel is an SPMD kernel or not.
 
  InsertPointTy createTargetInit(const LocationDescription &Loc, bool IsSPMD);
 
 
 
  /// Create a runtime call for kmpc_target_deinit
 
  ///
 
  /// \param Loc The insert and source location description.
 
  /// \param IsSPMD Flag to indicate if the kernel is an SPMD kernel or not.
 
  void createTargetDeinit(const LocationDescription &Loc, bool IsSPMD);
 
 
 
  ///}
 
 
 
private:
 
  // Sets the function attributes expected for the outlined function
 
  void setOutlinedTargetRegionFunctionAttributes(Function *OutlinedFn,
 
                                                 int32_t NumTeams,
 
                                                 int32_t NumThreads);
 
 
 
  // Creates the function ID/Address for the given outlined function.
 
  // In the case of an embedded device function the address of the function is
 
  // used, in the case of a non-offload function a constant is created.
 
  Constant *createOutlinedFunctionID(Function *OutlinedFn,
 
                                     StringRef EntryFnIDName);
 
 
 
  // Creates the region entry address for the outlined function
 
  Constant *createTargetRegionEntryAddr(Function *OutlinedFunction,
 
                                        StringRef EntryFnName);
 
 
 
public:
 
  /// Functions used to generate a function with the given name.
 
  using FunctionGenCallback = std::function<Function *(StringRef FunctionName)>;
 
 
 
  /// Create a unique name for the entry function using the source location
 
  /// information of the current target region. The name will be something like:
 
  ///
 
  /// __omp_offloading_DD_FFFF_PP_lBB[_CC]
 
  ///
 
  /// where DD_FFFF is an ID unique to the file (device and file IDs), PP is the
 
  /// mangled name of the function that encloses the target region and BB is the
 
  /// line number of the target region. CC is a count added when more than one
 
  /// region is located at the same location.
 
  ///
 
  /// If this target outline function is not an offload entry, we don't need to
 
  /// register it. This may happen if it is guarded by an if clause that is
 
  /// false at compile time, or no target archs have been specified.
 
  ///
 
  /// The created target region ID is used by the runtime library to identify
 
  /// the current target region, so it only has to be unique and not
 
  /// necessarily point to anything. It could be the pointer to the outlined
 
  /// function that implements the target region, but we aren't using that so
 
  /// that the compiler doesn't need to keep that, and could therefore inline
 
  /// the host function if proven worthwhile during optimization. In the other
 
  /// hand, if emitting code for the device, the ID has to be the function
 
  /// address so that it can retrieved from the offloading entry and launched
 
  /// by the runtime library. We also mark the outlined function to have
 
  /// external linkage in case we are emitting code for the device, because
 
  /// these functions will be entry points to the device.
 
  ///
 
  /// \param InfoManager The info manager keeping track of the offload entries
 
  /// \param EntryInfo The entry information about the function
 
  /// \param GenerateFunctionCallback The callback function to generate the code
 
  /// \param NumTeams Number default teams
 
  /// \param NumThreads Number default threads
 
  /// \param OutlinedFunction Pointer to the outlined function
 
  /// \param EntryFnIDName Name of the ID o be created
 
  void emitTargetRegionFunction(OffloadEntriesInfoManager &InfoManager,
 
                                TargetRegionEntryInfo &EntryInfo,
 
                                FunctionGenCallback &GenerateFunctionCallback,
 
                                int32_t NumTeams, int32_t NumThreads,
 
                                bool IsOffloadEntry, Function *&OutlinedFn,
 
                                Constant *&OutlinedFnID);
 
 
 
  /// Registers the given function and sets up the attribtues of the function
 
  /// Returns the FunctionID.
 
  ///
 
  /// \param InfoManager The info manager keeping track of the offload entries
 
  /// \param EntryInfo The entry information about the function
 
  /// \param OutlinedFunction Pointer to the outlined function
 
  /// \param EntryFnName Name of the outlined function
 
  /// \param EntryFnIDName Name of the ID o be created
 
  /// \param NumTeams Number default teams
 
  /// \param NumThreads Number default threads
 
  Constant *registerTargetRegionFunction(OffloadEntriesInfoManager &InfoManager,
 
                                         TargetRegionEntryInfo &EntryInfo,
 
                                         Function *OutlinedFunction,
 
                                         StringRef EntryFnName,
 
                                         StringRef EntryFnIDName,
 
                                         int32_t NumTeams, int32_t NumThreads);
 
 
 
  /// Declarations for LLVM-IR types (simple, array, function and structure) are
 
  /// generated below. Their names are defined and used in OpenMPKinds.def. Here
 
  /// we provide the declarations, the initializeTypes function will provide the
 
  /// values.
 
  ///
 
  ///{
 
#define OMP_TYPE(VarName, InitValue) Type *VarName = nullptr;
 
#define OMP_ARRAY_TYPE(VarName, ElemTy, ArraySize)                             \
 
  ArrayType *VarName##Ty = nullptr;                                            \
 
  PointerType *VarName##PtrTy = nullptr;
 
#define OMP_FUNCTION_TYPE(VarName, IsVarArg, ReturnType, ...)                  \
 
  FunctionType *VarName = nullptr;                                             \
 
  PointerType *VarName##Ptr = nullptr;
 
#define OMP_STRUCT_TYPE(VarName, StrName, ...)                                 \
 
  StructType *VarName = nullptr;                                               \
 
  PointerType *VarName##Ptr = nullptr;
 
#include "llvm/Frontend/OpenMP/OMPKinds.def"
 
 
 
  ///}
 
 
 
private:
 
  /// Create all simple and struct types exposed by the runtime and remember
 
  /// the llvm::PointerTypes of them for easy access later.
 
  void initializeTypes(Module &M);
 
 
 
  /// Common interface for generating entry calls for OMP Directives.
 
  /// if the directive has a region/body, It will set the insertion
 
  /// point to the body
 
  ///
 
  /// \param OMPD Directive to generate entry blocks for
 
  /// \param EntryCall Call to the entry OMP Runtime Function
 
  /// \param ExitBB block where the region ends.
 
  /// \param Conditional indicate if the entry call result will be used
 
  ///        to evaluate a conditional of whether a thread will execute
 
  ///        body code or not.
 
  ///
 
  /// \return The insertion position in exit block
 
  InsertPointTy emitCommonDirectiveEntry(omp::Directive OMPD, Value *EntryCall,
 
                                         BasicBlock *ExitBB,
 
                                         bool Conditional = false);
 
 
 
  /// Common interface to finalize the region
 
  ///
 
  /// \param OMPD Directive to generate exiting code for
 
  /// \param FinIP Insertion point for emitting Finalization code and exit call
 
  /// \param ExitCall Call to the ending OMP Runtime Function
 
  /// \param HasFinalize indicate if the directive will require finalization
 
  ///         and has a finalization callback in the stack that
 
  ///        should be called.
 
  ///
 
  /// \return The insertion position in exit block
 
  InsertPointTy emitCommonDirectiveExit(omp::Directive OMPD,
 
                                        InsertPointTy FinIP,
 
                                        Instruction *ExitCall,
 
                                        bool HasFinalize = true);
 
 
 
  /// Common Interface to generate OMP inlined regions
 
  ///
 
  /// \param OMPD Directive to generate inlined region for
 
  /// \param EntryCall Call to the entry OMP Runtime Function
 
  /// \param ExitCall Call to the ending OMP Runtime Function
 
  /// \param BodyGenCB Body code generation callback.
 
  /// \param FiniCB Finalization Callback. Will be called when finalizing region
 
  /// \param Conditional indicate if the entry call result will be used
 
  ///        to evaluate a conditional of whether a thread will execute
 
  ///        body code or not.
 
  /// \param HasFinalize indicate if the directive will require finalization
 
  ///        and has a finalization callback in the stack that
 
  ///        should be called.
 
  /// \param IsCancellable if HasFinalize is set to true, indicate if the
 
  ///        the directive should be cancellable.
 
  /// \return The insertion point after the region
 
 
 
  InsertPointTy
 
  EmitOMPInlinedRegion(omp::Directive OMPD, Instruction *EntryCall,
 
                       Instruction *ExitCall, BodyGenCallbackTy BodyGenCB,
 
                       FinalizeCallbackTy FiniCB, bool Conditional = false,
 
                       bool HasFinalize = true, bool IsCancellable = false);
 
 
 
  /// Get the platform-specific name separator.
 
  /// \param Parts different parts of the final name that needs separation
 
  /// \param FirstSeparator First separator used between the initial two
 
  ///        parts of the name.
 
  /// \param Separator separator used between all of the rest consecutive
 
  ///        parts of the name
 
  static std::string getNameWithSeparators(ArrayRef<StringRef> Parts,
 
                                           StringRef FirstSeparator,
 
                                           StringRef Separator);
 
 
 
  /// Returns corresponding lock object for the specified critical region
 
  /// name. If the lock object does not exist it is created, otherwise the
 
  /// reference to the existing copy is returned.
 
  /// \param CriticalName Name of the critical region.
 
  ///
 
  Value *getOMPCriticalRegionLock(StringRef CriticalName);
 
 
 
  /// Callback type for Atomic Expression update
 
  /// ex:
 
  /// \code{.cpp}
 
  /// unsigned x = 0;
 
  /// #pragma omp atomic update
 
  /// x = Expr(x_old);  //Expr() is any legal operation
 
  /// \endcode
 
  ///
 
  /// \param XOld the value of the atomic memory address to use for update
 
  /// \param IRB reference to the IRBuilder to use
 
  ///
 
  /// \returns Value to update X to.
 
  using AtomicUpdateCallbackTy =
 
      const function_ref<Value *(Value *XOld, IRBuilder<> &IRB)>;
 
 
 
private:
 
  enum AtomicKind { Read, Write, Update, Capture, Compare };
 
 
 
  /// Determine whether to emit flush or not
 
  ///
 
  /// \param Loc    The insert and source location description.
 
  /// \param AO     The required atomic ordering
 
  /// \param AK     The OpenMP atomic operation kind used.
 
  ///
 
  /// \returns          wether a flush was emitted or not
 
  bool checkAndEmitFlushAfterAtomic(const LocationDescription &Loc,
 
                                    AtomicOrdering AO, AtomicKind AK);
 
 
 
  /// Emit atomic update for constructs: X = X BinOp Expr ,or X = Expr BinOp X
 
  /// For complex Operations: X = UpdateOp(X) => CmpExch X, old_X, UpdateOp(X)
 
  /// Only Scalar data types.
 
  ///
 
  /// \param AllocaIP     The insertion point to be used for alloca
 
  ///                   instructions.
 
  /// \param X                      The target atomic pointer to be updated
 
  /// \param XElemTy    The element type of the atomic pointer.
 
  /// \param Expr                   The value to update X with.
 
  /// \param AO                     Atomic ordering of the generated atomic
 
  ///                   instructions.
 
  /// \param RMWOp                The binary operation used for update. If
 
  ///                   operation is not supported by atomicRMW,
 
  ///                   or belong to {FADD, FSUB, BAD_BINOP}.
 
  ///                   Then a `cmpExch` based  atomic will be generated.
 
  /// \param UpdateOp   Code generator for complex expressions that cannot be
 
  ///                   expressed through atomicrmw instruction.
 
  /// \param VolatileX       true if \a X volatile?
 
  /// \param IsXBinopExpr true if \a X is Left H.S. in Right H.S. part of the
 
  ///                     update expression, false otherwise.
 
  ///                     (e.g. true for X = X BinOp Expr)
 
  ///
 
  /// \returns A pair of the old value of X before the update, and the value
 
  ///          used for the update.
 
  std::pair<Value *, Value *>
 
  emitAtomicUpdate(InsertPointTy AllocaIP, Value *X, Type *XElemTy, Value *Expr,
 
                   AtomicOrdering AO, AtomicRMWInst::BinOp RMWOp,
 
                   AtomicUpdateCallbackTy &UpdateOp, bool VolatileX,
 
                   bool IsXBinopExpr);
 
 
 
  /// Emit the binary op. described by \p RMWOp, using \p Src1 and \p Src2 .
 
  ///
 
  /// \Return The instruction
 
  Value *emitRMWOpAsInstruction(Value *Src1, Value *Src2,
 
                                AtomicRMWInst::BinOp RMWOp);
 
 
 
public:
 
  /// a struct to pack relevant information while generating atomic Ops
 
  struct AtomicOpValue {
 
    Value *Var = nullptr;
 
    Type *ElemTy = nullptr;
 
    bool IsSigned = false;
 
    bool IsVolatile = false;
 
  };
 
 
 
  /// Emit atomic Read for : V = X --- Only Scalar data types.
 
  ///
 
  /// \param Loc    The insert and source location description.
 
  /// \param X                  The target pointer to be atomically read
 
  /// \param V                  Memory address where to store atomically read
 
  ///                                       value
 
  /// \param AO                 Atomic ordering of the generated atomic
 
  ///                                       instructions.
 
  ///
 
  /// \return Insertion point after generated atomic read IR.
 
  InsertPointTy createAtomicRead(const LocationDescription &Loc,
 
                                 AtomicOpValue &X, AtomicOpValue &V,
 
                                 AtomicOrdering AO);
 
 
 
  /// Emit atomic write for : X = Expr --- Only Scalar data types.
 
  ///
 
  /// \param Loc    The insert and source location description.
 
  /// \param X                  The target pointer to be atomically written to
 
  /// \param Expr               The value to store.
 
  /// \param AO                 Atomic ordering of the generated atomic
 
  ///               instructions.
 
  ///
 
  /// \return Insertion point after generated atomic Write IR.
 
  InsertPointTy createAtomicWrite(const LocationDescription &Loc,
 
                                  AtomicOpValue &X, Value *Expr,
 
                                  AtomicOrdering AO);
 
 
 
  /// Emit atomic update for constructs: X = X BinOp Expr ,or X = Expr BinOp X
 
  /// For complex Operations: X = UpdateOp(X) => CmpExch X, old_X, UpdateOp(X)
 
  /// Only Scalar data types.
 
  ///
 
  /// \param Loc      The insert and source location description.
 
  /// \param AllocaIP The insertion point to be used for alloca instructions.
 
  /// \param X        The target atomic pointer to be updated
 
  /// \param Expr     The value to update X with.
 
  /// \param AO       Atomic ordering of the generated atomic instructions.
 
  /// \param RMWOp    The binary operation used for update. If operation
 
  ///                 is        not supported by atomicRMW, or belong to
 
  ///                   {FADD, FSUB, BAD_BINOP}. Then a `cmpExch` based
 
  ///                 atomic will be generated.
 
  /// \param UpdateOp   Code generator for complex expressions that cannot be
 
  ///                   expressed through atomicrmw instruction.
 
  /// \param IsXBinopExpr true if \a X is Left H.S. in Right H.S. part of the
 
  ///                     update expression, false otherwise.
 
  ///                       (e.g. true for X = X BinOp Expr)
 
  ///
 
  /// \return Insertion point after generated atomic update IR.
 
  InsertPointTy createAtomicUpdate(const LocationDescription &Loc,
 
                                   InsertPointTy AllocaIP, AtomicOpValue &X,
 
                                   Value *Expr, AtomicOrdering AO,
 
                                   AtomicRMWInst::BinOp RMWOp,
 
                                   AtomicUpdateCallbackTy &UpdateOp,
 
                                   bool IsXBinopExpr);
 
 
 
  /// Emit atomic update for constructs: --- Only Scalar data types
 
  /// V = X; X = X BinOp Expr ,
 
  /// X = X BinOp Expr; V = X,
 
  /// V = X; X = Expr BinOp X,
 
  /// X = Expr BinOp X; V = X,
 
  /// V = X; X = UpdateOp(X),
 
  /// X = UpdateOp(X); V = X,
 
  ///
 
  /// \param Loc        The insert and source location description.
 
  /// \param AllocaIP   The insertion point to be used for alloca instructions.
 
  /// \param X          The target atomic pointer to be updated
 
  /// \param V          Memory address where to store captured value
 
  /// \param Expr       The value to update X with.
 
  /// \param AO         Atomic ordering of the generated atomic instructions
 
  /// \param RMWOp      The binary operation used for update. If
 
  ///                   operation is not supported by atomicRMW, or belong to
 
  ///                     {FADD, FSUB, BAD_BINOP}. Then a cmpExch based
 
  ///                   atomic will be generated.
 
  /// \param UpdateOp   Code generator for complex expressions that cannot be
 
  ///                   expressed through atomicrmw instruction.
 
  /// \param UpdateExpr true if X is an in place update of the form
 
  ///                   X = X BinOp Expr or X = Expr BinOp X
 
  /// \param IsXBinopExpr true if X is Left H.S. in Right H.S. part of the
 
  ///                     update expression, false otherwise.
 
  ///                     (e.g. true for X = X BinOp Expr)
 
  /// \param IsPostfixUpdate true if original value of 'x' must be stored in
 
  ///                        'v', not an updated one.
 
  ///
 
  /// \return Insertion point after generated atomic capture IR.
 
  InsertPointTy
 
  createAtomicCapture(const LocationDescription &Loc, InsertPointTy AllocaIP,
 
                      AtomicOpValue &X, AtomicOpValue &V, Value *Expr,
 
                      AtomicOrdering AO, AtomicRMWInst::BinOp RMWOp,
 
                      AtomicUpdateCallbackTy &UpdateOp, bool UpdateExpr,
 
                      bool IsPostfixUpdate, bool IsXBinopExpr);
 
 
 
  /// Emit atomic compare for constructs: --- Only scalar data types
 
  /// cond-expr-stmt:
 
  /// x = x ordop expr ? expr : x;
 
  /// x = expr ordop x ? expr : x;
 
  /// x = x == e ? d : x;
 
  /// x = e == x ? d : x; (this one is not in the spec)
 
  /// cond-update-stmt:
 
  /// if (x ordop expr) { x = expr; }
 
  /// if (expr ordop x) { x = expr; }
 
  /// if (x == e) { x = d; }
 
  /// if (e == x) { x = d; } (this one is not in the spec)
 
  /// conditional-update-capture-atomic:
 
  /// v = x; cond-update-stmt; (IsPostfixUpdate=true, IsFailOnly=false)
 
  /// cond-update-stmt; v = x; (IsPostfixUpdate=false, IsFailOnly=false)
 
  /// if (x == e) { x = d; } else { v = x; } (IsPostfixUpdate=false,
 
  ///                                         IsFailOnly=true)
 
  /// r = x == e; if (r) { x = d; } (IsPostfixUpdate=false, IsFailOnly=false)
 
  /// r = x == e; if (r) { x = d; } else { v = x; } (IsPostfixUpdate=false,
 
  ///                                                IsFailOnly=true)
 
  ///
 
  /// \param Loc          The insert and source location description.
 
  /// \param X            The target atomic pointer to be updated.
 
  /// \param V            Memory address where to store captured value (for
 
  ///                     compare capture only).
 
  /// \param R            Memory address where to store comparison result
 
  ///                     (for compare capture with '==' only).
 
  /// \param E            The expected value ('e') for forms that use an
 
  ///                     equality comparison or an expression ('expr') for
 
  ///                     forms that use 'ordop' (logically an atomic maximum or
 
  ///                     minimum).
 
  /// \param D            The desired value for forms that use an equality
 
  ///                     comparison. If forms that use 'ordop', it should be
 
  ///                     \p nullptr.
 
  /// \param AO           Atomic ordering of the generated atomic instructions.
 
  /// \param Op           Atomic compare operation. It can only be ==, <, or >.
 
  /// \param IsXBinopExpr True if the conditional statement is in the form where
 
  ///                     x is on LHS. It only matters for < or >.
 
  /// \param IsPostfixUpdate  True if original value of 'x' must be stored in
 
  ///                         'v', not an updated one (for compare capture
 
  ///                         only).
 
  /// \param IsFailOnly   True if the original value of 'x' is stored to 'v'
 
  ///                     only when the comparison fails. This is only valid for
 
  ///                     the case the comparison is '=='.
 
  ///
 
  /// \return Insertion point after generated atomic capture IR.
 
  InsertPointTy
 
  createAtomicCompare(const LocationDescription &Loc, AtomicOpValue &X,
 
                      AtomicOpValue &V, AtomicOpValue &R, Value *E, Value *D,
 
                      AtomicOrdering AO, omp::OMPAtomicCompareOp Op,
 
                      bool IsXBinopExpr, bool IsPostfixUpdate, bool IsFailOnly);
 
 
 
  /// Create the control flow structure of a canonical OpenMP loop.
 
  ///
 
  /// The emitted loop will be disconnected, i.e. no edge to the loop's
 
  /// preheader and no terminator in the AfterBB. The OpenMPIRBuilder's
 
  /// IRBuilder location is not preserved.
 
  ///
 
  /// \param DL        DebugLoc used for the instructions in the skeleton.
 
  /// \param TripCount Value to be used for the trip count.
 
  /// \param F         Function in which to insert the BasicBlocks.
 
  /// \param PreInsertBefore  Where to insert BBs that execute before the body,
 
  ///                         typically the body itself.
 
  /// \param PostInsertBefore Where to insert BBs that execute after the body.
 
  /// \param Name      Base name used to derive BB
 
  ///                  and instruction names.
 
  ///
 
  /// \returns The CanonicalLoopInfo that represents the emitted loop.
 
  CanonicalLoopInfo *createLoopSkeleton(DebugLoc DL, Value *TripCount,
 
                                        Function *F,
 
                                        BasicBlock *PreInsertBefore,
 
                                        BasicBlock *PostInsertBefore,
 
                                        const Twine &Name = {});
 
  /// OMP Offload Info Metadata name string
 
  const std::string ompOffloadInfoName = "omp_offload.info";
 
 
 
  /// Loads all the offload entries information from the host IR
 
  /// metadata. This function is only meant to be used with device code
 
  /// generation.
 
  ///
 
  /// \param M         Module to load Metadata info from. Module passed maybe
 
  /// loaded from bitcode file, i.e, different from OpenMPIRBuilder::M module.
 
  /// \param OffloadEntriesInfoManager Initialize Offload Entry information.
 
  void
 
  loadOffloadInfoMetadata(Module &M,
 
                          OffloadEntriesInfoManager &OffloadEntriesInfoManager);
 
 
 
  /// Gets (if variable with the given name already exist) or creates
 
  /// internal global variable with the specified Name. The created variable has
 
  /// linkage CommonLinkage by default and is initialized by null value.
 
  /// \param Ty Type of the global variable. If it is exist already the type
 
  /// must be the same.
 
  /// \param Name Name of the variable.
 
  GlobalVariable *getOrCreateInternalVariable(Type *Ty, const StringRef &Name,
 
                                              unsigned AddressSpace = 0);
 
};
 
 
 
/// Data structure to contain the information needed to uniquely identify
 
/// a target entry.
 
struct TargetRegionEntryInfo {
 
  std::string ParentName;
 
  unsigned DeviceID;
 
  unsigned FileID;
 
  unsigned Line;
 
  unsigned Count;
 
 
 
  TargetRegionEntryInfo()
 
      : ParentName(""), DeviceID(0), FileID(0), Line(0), Count(0) {}
 
  TargetRegionEntryInfo(StringRef ParentName, unsigned DeviceID,
 
                        unsigned FileID, unsigned Line, unsigned Count = 0)
 
      : ParentName(ParentName), DeviceID(DeviceID), FileID(FileID), Line(Line),
 
        Count(Count) {}
 
 
 
  static void getTargetRegionEntryFnName(SmallVectorImpl<char> &Name,
 
                                         StringRef ParentName,
 
                                         unsigned DeviceID, unsigned FileID,
 
                                         unsigned Line, unsigned Count);
 
 
 
  bool operator<(const TargetRegionEntryInfo RHS) const {
 
    return std::make_tuple(ParentName, DeviceID, FileID, Line, Count) <
 
           std::make_tuple(RHS.ParentName, RHS.DeviceID, RHS.FileID, RHS.Line,
 
                           RHS.Count);
 
  }
 
};
 
 
 
/// Class that manages information about offload code regions and data
 
class OffloadEntriesInfoManager {
 
  /// Number of entries registered so far.
 
  OpenMPIRBuilderConfig Config;
 
  unsigned OffloadingEntriesNum = 0;
 
 
 
public:
 
  void setConfig(OpenMPIRBuilderConfig C) { Config = C; }
 
 
 
  /// Base class of the entries info.
 
  class OffloadEntryInfo {
 
  public:
 
    /// Kind of a given entry.
 
    enum OffloadingEntryInfoKinds : unsigned {
 
      /// Entry is a target region.
 
      OffloadingEntryInfoTargetRegion = 0,
 
      /// Entry is a declare target variable.
 
      OffloadingEntryInfoDeviceGlobalVar = 1,
 
      /// Invalid entry info.
 
      OffloadingEntryInfoInvalid = ~0u
 
    };
 
 
 
  protected:
 
    OffloadEntryInfo() = delete;
 
    explicit OffloadEntryInfo(OffloadingEntryInfoKinds Kind) : Kind(Kind) {}
 
    explicit OffloadEntryInfo(OffloadingEntryInfoKinds Kind, unsigned Order,
 
                              uint32_t Flags)
 
        : Flags(Flags), Order(Order), Kind(Kind) {}
 
    ~OffloadEntryInfo() = default;
 
 
 
  public:
 
    bool isValid() const { return Order != ~0u; }
 
    unsigned getOrder() const { return Order; }
 
    OffloadingEntryInfoKinds getKind() const { return Kind; }
 
    uint32_t getFlags() const { return Flags; }
 
    void setFlags(uint32_t NewFlags) { Flags = NewFlags; }
 
    Constant *getAddress() const { return cast_or_null<Constant>(Addr); }
 
    void setAddress(Constant *V) {
 
      assert(!Addr.pointsToAliveValue() && "Address has been set before!");
 
      Addr = V;
 
    }
 
    static bool classof(const OffloadEntryInfo *Info) { return true; }
 
 
 
  private:
 
    /// Address of the entity that has to be mapped for offloading.
 
    WeakTrackingVH Addr;
 
 
 
    /// Flags associated with the device global.
 
    uint32_t Flags = 0u;
 
 
 
    /// Order this entry was emitted.
 
    unsigned Order = ~0u;
 
 
 
    OffloadingEntryInfoKinds Kind = OffloadingEntryInfoInvalid;
 
  };
 
 
 
  /// Return true if a there are no entries defined.
 
  bool empty() const;
 
  /// Return number of entries defined so far.
 
  unsigned size() const { return OffloadingEntriesNum; }
 
 
 
  OffloadEntriesInfoManager() : Config() {}
 
 
 
  //
 
  // Target region entries related.
 
  //
 
 
 
  /// Kind of the target registry entry.
 
  enum OMPTargetRegionEntryKind : uint32_t {
 
    /// Mark the entry as target region.
 
    OMPTargetRegionEntryTargetRegion = 0x0,
 
    /// Mark the entry as a global constructor.
 
    OMPTargetRegionEntryCtor = 0x02,
 
    /// Mark the entry as a global destructor.
 
    OMPTargetRegionEntryDtor = 0x04,
 
  };
 
 
 
  /// Target region entries info.
 
  class OffloadEntryInfoTargetRegion final : public OffloadEntryInfo {
 
    /// Address that can be used as the ID of the entry.
 
    Constant *ID = nullptr;
 
 
 
  public:
 
    OffloadEntryInfoTargetRegion()
 
        : OffloadEntryInfo(OffloadingEntryInfoTargetRegion) {}
 
    explicit OffloadEntryInfoTargetRegion(unsigned Order, Constant *Addr,
 
                                          Constant *ID,
 
                                          OMPTargetRegionEntryKind Flags)
 
        : OffloadEntryInfo(OffloadingEntryInfoTargetRegion, Order, Flags),
 
          ID(ID) {
 
      setAddress(Addr);
 
    }
 
 
 
    Constant *getID() const { return ID; }
 
    void setID(Constant *V) {
 
      assert(!ID && "ID has been set before!");
 
      ID = V;
 
    }
 
    static bool classof(const OffloadEntryInfo *Info) {
 
      return Info->getKind() == OffloadingEntryInfoTargetRegion;
 
    }
 
  };
 
 
 
  /// Initialize target region entry.
 
  /// This is ONLY needed for DEVICE compilation.
 
  void initializeTargetRegionEntryInfo(const TargetRegionEntryInfo &EntryInfo,
 
                                       unsigned Order);
 
  /// Register target region entry.
 
  void registerTargetRegionEntryInfo(TargetRegionEntryInfo EntryInfo,
 
                                     Constant *Addr, Constant *ID,
 
                                     OMPTargetRegionEntryKind Flags);
 
  /// Return true if a target region entry with the provided information
 
  /// exists.
 
  bool hasTargetRegionEntryInfo(TargetRegionEntryInfo EntryInfo,
 
                                bool IgnoreAddressId = false) const;
 
 
 
  // Return the Name based on \a EntryInfo using the next available Count.
 
  void getTargetRegionEntryFnName(SmallVectorImpl<char> &Name,
 
                                  const TargetRegionEntryInfo &EntryInfo);
 
 
 
  /// brief Applies action \a Action on all registered entries.
 
  typedef function_ref<void(const TargetRegionEntryInfo &EntryInfo,
 
                            const OffloadEntryInfoTargetRegion &)>
 
      OffloadTargetRegionEntryInfoActTy;
 
  void
 
  actOnTargetRegionEntriesInfo(const OffloadTargetRegionEntryInfoActTy &Action);
 
 
 
  //
 
  // Device global variable entries related.
 
  //
 
 
 
  /// Kind of the global variable entry..
 
  enum OMPTargetGlobalVarEntryKind : uint32_t {
 
    /// Mark the entry as a to declare target.
 
    OMPTargetGlobalVarEntryTo = 0x0,
 
    /// Mark the entry as a to declare target link.
 
    OMPTargetGlobalVarEntryLink = 0x1,
 
  };
 
 
 
  /// Device global variable entries info.
 
  class OffloadEntryInfoDeviceGlobalVar final : public OffloadEntryInfo {
 
    /// Type of the global variable.
 
    int64_t VarSize;
 
    GlobalValue::LinkageTypes Linkage;
 
 
 
  public:
 
    OffloadEntryInfoDeviceGlobalVar()
 
        : OffloadEntryInfo(OffloadingEntryInfoDeviceGlobalVar) {}
 
    explicit OffloadEntryInfoDeviceGlobalVar(unsigned Order,
 
                                             OMPTargetGlobalVarEntryKind Flags)
 
        : OffloadEntryInfo(OffloadingEntryInfoDeviceGlobalVar, Order, Flags) {}
 
    explicit OffloadEntryInfoDeviceGlobalVar(unsigned Order, Constant *Addr,
 
                                             int64_t VarSize,
 
                                             OMPTargetGlobalVarEntryKind Flags,
 
                                             GlobalValue::LinkageTypes Linkage)
 
        : OffloadEntryInfo(OffloadingEntryInfoDeviceGlobalVar, Order, Flags),
 
          VarSize(VarSize), Linkage(Linkage) {
 
      setAddress(Addr);
 
    }
 
 
 
    int64_t getVarSize() const { return VarSize; }
 
    void setVarSize(int64_t Size) { VarSize = Size; }
 
    GlobalValue::LinkageTypes getLinkage() const { return Linkage; }
 
    void setLinkage(GlobalValue::LinkageTypes LT) { Linkage = LT; }
 
    static bool classof(const OffloadEntryInfo *Info) {
 
      return Info->getKind() == OffloadingEntryInfoDeviceGlobalVar;
 
    }
 
  };
 
 
 
  /// Initialize device global variable entry.
 
  /// This is ONLY used for DEVICE compilation.
 
  void initializeDeviceGlobalVarEntryInfo(StringRef Name,
 
                                          OMPTargetGlobalVarEntryKind Flags,
 
                                          unsigned Order);
 
 
 
  /// Register device global variable entry.
 
  void registerDeviceGlobalVarEntryInfo(StringRef VarName, Constant *Addr,
 
                                        int64_t VarSize,
 
                                        OMPTargetGlobalVarEntryKind Flags,
 
                                        GlobalValue::LinkageTypes Linkage);
 
  /// Checks if the variable with the given name has been registered already.
 
  bool hasDeviceGlobalVarEntryInfo(StringRef VarName) const {
 
    return OffloadEntriesDeviceGlobalVar.count(VarName) > 0;
 
  }
 
  /// Applies action \a Action on all registered entries.
 
  typedef function_ref<void(StringRef, const OffloadEntryInfoDeviceGlobalVar &)>
 
      OffloadDeviceGlobalVarEntryInfoActTy;
 
  void actOnDeviceGlobalVarEntriesInfo(
 
      const OffloadDeviceGlobalVarEntryInfoActTy &Action);
 
 
 
private:
 
  /// Return the count of entries at a particular source location.
 
  unsigned
 
  getTargetRegionEntryInfoCount(const TargetRegionEntryInfo &EntryInfo) const;
 
 
 
  /// Update the count of entries at a particular source location.
 
  void
 
  incrementTargetRegionEntryInfoCount(const TargetRegionEntryInfo &EntryInfo);
 
 
 
  static TargetRegionEntryInfo
 
  getTargetRegionEntryCountKey(const TargetRegionEntryInfo &EntryInfo) {
 
    return TargetRegionEntryInfo(EntryInfo.ParentName, EntryInfo.DeviceID,
 
                                 EntryInfo.FileID, EntryInfo.Line, 0);
 
  }
 
 
 
  // Count of entries at a location.
 
  std::map<TargetRegionEntryInfo, unsigned> OffloadEntriesTargetRegionCount;
 
 
 
  // Storage for target region entries kind.
 
  typedef std::map<TargetRegionEntryInfo, OffloadEntryInfoTargetRegion>
 
      OffloadEntriesTargetRegionTy;
 
  OffloadEntriesTargetRegionTy OffloadEntriesTargetRegion;
 
  /// Storage for device global variable entries kind. The storage is to be
 
  /// indexed by mangled name.
 
  typedef StringMap<OffloadEntryInfoDeviceGlobalVar>
 
      OffloadEntriesDeviceGlobalVarTy;
 
  OffloadEntriesDeviceGlobalVarTy OffloadEntriesDeviceGlobalVar;
 
};
 
 
 
/// Class to represented the control flow structure of an OpenMP canonical loop.
 
///
 
/// The control-flow structure is standardized for easy consumption by
 
/// directives associated with loops. For instance, the worksharing-loop
 
/// construct may change this control flow such that each loop iteration is
 
/// executed on only one thread. The constraints of a canonical loop in brief
 
/// are:
 
///
 
///  * The number of loop iterations must have been computed before entering the
 
///    loop.
 
///
 
///  * Has an (unsigned) logical induction variable that starts at zero and
 
///    increments by one.
 
///
 
///  * The loop's CFG itself has no side-effects. The OpenMP specification
 
///    itself allows side-effects, but the order in which they happen, including
 
///    how often or whether at all, is unspecified. We expect that the frontend
 
///    will emit those side-effect instructions somewhere (e.g. before the loop)
 
///    such that the CanonicalLoopInfo itself can be side-effect free.
 
///
 
/// Keep in mind that CanonicalLoopInfo is meant to only describe a repeated
 
/// execution of a loop body that satifies these constraints. It does NOT
 
/// represent arbitrary SESE regions that happen to contain a loop. Do not use
 
/// CanonicalLoopInfo for such purposes.
 
///
 
/// The control flow can be described as follows:
 
///
 
///     Preheader
 
///        |
 
///  /-> Header
 
///  |     |
 
///  |    Cond---\
 
///  |     |     |
 
///  |    Body   |
 
///  |    | |    |
 
///  |   <...>   |
 
///  |    | |    |
 
///   \--Latch   |
 
///              |
 
///             Exit
 
///              |
 
///            After
 
///
 
/// The loop is thought to start at PreheaderIP (at the Preheader's terminator,
 
/// including) and end at AfterIP (at the After's first instruction, excluding).
 
/// That is, instructions in the Preheader and After blocks (except the
 
/// Preheader's terminator) are out of CanonicalLoopInfo's control and may have
 
/// side-effects. Typically, the Preheader is used to compute the loop's trip
 
/// count. The instructions from BodyIP (at the Body block's first instruction,
 
/// excluding) until the Latch are also considered outside CanonicalLoopInfo's
 
/// control and thus can have side-effects. The body block is the single entry
 
/// point into the loop body, which may contain arbitrary control flow as long
 
/// as all control paths eventually branch to the Latch block.
 
///
 
/// TODO: Consider adding another standardized BasicBlock between Body CFG and
 
/// Latch to guarantee that there is only a single edge to the latch. It would
 
/// make loop transformations easier to not needing to consider multiple
 
/// predecessors of the latch (See redirectAllPredecessorsTo) and would give us
 
/// an equivalant to PreheaderIP, AfterIP and BodyIP for inserting code that
 
/// executes after each body iteration.
 
///
 
/// There must be no loop-carried dependencies through llvm::Values. This is
 
/// equivalant to that the Latch has no PHINode and the Header's only PHINode is
 
/// for the induction variable.
 
///
 
/// All code in Header, Cond, Latch and Exit (plus the terminator of the
 
/// Preheader) are CanonicalLoopInfo's responsibility and their build-up checked
 
/// by assertOK(). They are expected to not be modified unless explicitly
 
/// modifying the CanonicalLoopInfo through a methods that applies a OpenMP
 
/// loop-associated construct such as applyWorkshareLoop, tileLoops, unrollLoop,
 
/// etc. These methods usually invalidate the CanonicalLoopInfo and re-use its
 
/// basic blocks. After invalidation, the CanonicalLoopInfo must not be used
 
/// anymore as its underlying control flow may not exist anymore.
 
/// Loop-transformation methods such as tileLoops, collapseLoops and unrollLoop
 
/// may also return a new CanonicalLoopInfo that can be passed to other
 
/// loop-associated construct implementing methods. These loop-transforming
 
/// methods may either create a new CanonicalLoopInfo usually using
 
/// createLoopSkeleton and invalidate the input CanonicalLoopInfo, or reuse and
 
/// modify one of the input CanonicalLoopInfo and return it as representing the
 
/// modified loop. What is done is an implementation detail of
 
/// transformation-implementing method and callers should always assume that the
 
/// CanonicalLoopInfo passed to it is invalidated and a new object is returned.
 
/// Returned CanonicalLoopInfo have the same structure and guarantees as the one
 
/// created by createCanonicalLoop, such that transforming methods do not have
 
/// to special case where the CanonicalLoopInfo originated from.
 
///
 
/// Generally, methods consuming CanonicalLoopInfo do not need an
 
/// OpenMPIRBuilder::InsertPointTy as argument, but use the locations of the
 
/// CanonicalLoopInfo to insert new or modify existing instructions. Unless
 
/// documented otherwise, methods consuming CanonicalLoopInfo do not invalidate
 
/// any InsertPoint that is outside CanonicalLoopInfo's control. Specifically,
 
/// any InsertPoint in the Preheader, After or Block can still be used after
 
/// calling such a method.
 
///
 
/// TODO: Provide mechanisms for exception handling and cancellation points.
 
///
 
/// Defined outside OpenMPIRBuilder because nested classes cannot be
 
/// forward-declared, e.g. to avoid having to include the entire OMPIRBuilder.h.
 
class CanonicalLoopInfo {
 
  friend class OpenMPIRBuilder;
 
 
 
private:
 
  BasicBlock *Header = nullptr;
 
  BasicBlock *Cond = nullptr;
 
  BasicBlock *Latch = nullptr;
 
  BasicBlock *Exit = nullptr;
 
 
 
  /// Add the control blocks of this loop to \p BBs.
 
  ///
 
  /// This does not include any block from the body, including the one returned
 
  /// by getBody().
 
  ///
 
  /// FIXME: This currently includes the Preheader and After blocks even though
 
  /// their content is (mostly) not under CanonicalLoopInfo's control.
 
  /// Re-evaluated whether this makes sense.
 
  void collectControlBlocks(SmallVectorImpl<BasicBlock *> &BBs);
 
 
 
  /// Sets the number of loop iterations to the given value. This value must be
 
  /// valid in the condition block (i.e., defined in the preheader) and is
 
  /// interpreted as an unsigned integer.
 
  void setTripCount(Value *TripCount);
 
 
 
  /// Replace all uses of the canonical induction variable in the loop body with
 
  /// a new one.
 
  ///
 
  /// The intended use case is to update the induction variable for an updated
 
  /// iteration space such that it can stay normalized in the 0...tripcount-1
 
  /// range.
 
  ///
 
  /// The \p Updater is called with the (presumable updated) current normalized
 
  /// induction variable and is expected to return the value that uses of the
 
  /// pre-updated induction values should use instead, typically dependent on
 
  /// the new induction variable. This is a lambda (instead of e.g. just passing
 
  /// the new value) to be able to distinguish the uses of the pre-updated
 
  /// induction variable and uses of the induction varible to compute the
 
  /// updated induction variable value.
 
  void mapIndVar(llvm::function_ref<Value *(Instruction *)> Updater);
 
 
 
public:
 
  /// Returns whether this object currently represents the IR of a loop. If
 
  /// returning false, it may have been consumed by a loop transformation or not
 
  /// been intialized. Do not use in this case;
 
  bool isValid() const { return Header; }
 
 
 
  /// The preheader ensures that there is only a single edge entering the loop.
 
  /// Code that must be execute before any loop iteration can be emitted here,
 
  /// such as computing the loop trip count and begin lifetime markers. Code in
 
  /// the preheader is not considered part of the canonical loop.
 
  BasicBlock *getPreheader() const;
 
 
 
  /// The header is the entry for each iteration. In the canonical control flow,
 
  /// it only contains the PHINode for the induction variable.
 
  BasicBlock *getHeader() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    return Header;
 
  }
 
 
 
  /// The condition block computes whether there is another loop iteration. If
 
  /// yes, branches to the body; otherwise to the exit block.
 
  BasicBlock *getCond() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    return Cond;
 
  }
 
 
 
  /// The body block is the single entry for a loop iteration and not controlled
 
  /// by CanonicalLoopInfo. It can contain arbitrary control flow but must
 
  /// eventually branch to the \p Latch block.
 
  BasicBlock *getBody() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    return cast<BranchInst>(Cond->getTerminator())->getSuccessor(0);
 
  }
 
 
 
  /// Reaching the latch indicates the end of the loop body code. In the
 
  /// canonical control flow, it only contains the increment of the induction
 
  /// variable.
 
  BasicBlock *getLatch() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    return Latch;
 
  }
 
 
 
  /// Reaching the exit indicates no more iterations are being executed.
 
  BasicBlock *getExit() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    return Exit;
 
  }
 
 
 
  /// The after block is intended for clean-up code such as lifetime end
 
  /// markers. It is separate from the exit block to ensure, analogous to the
 
  /// preheader, it having just a single entry edge and being free from PHI
 
  /// nodes should there be multiple loop exits (such as from break
 
  /// statements/cancellations).
 
  BasicBlock *getAfter() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    return Exit->getSingleSuccessor();
 
  }
 
 
 
  /// Returns the llvm::Value containing the number of loop iterations. It must
 
  /// be valid in the preheader and always interpreted as an unsigned integer of
 
  /// any bit-width.
 
  Value *getTripCount() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    Instruction *CmpI = &Cond->front();
 
    assert(isa<CmpInst>(CmpI) && "First inst must compare IV with TripCount");
 
    return CmpI->getOperand(1);
 
  }
 
 
 
  /// Returns the instruction representing the current logical induction
 
  /// variable. Always unsigned, always starting at 0 with an increment of one.
 
  Instruction *getIndVar() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    Instruction *IndVarPHI = &Header->front();
 
    assert(isa<PHINode>(IndVarPHI) && "First inst must be the IV PHI");
 
    return IndVarPHI;
 
  }
 
 
 
  /// Return the type of the induction variable (and the trip count).
 
  Type *getIndVarType() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    return getIndVar()->getType();
 
  }
 
 
 
  /// Return the insertion point for user code before the loop.
 
  OpenMPIRBuilder::InsertPointTy getPreheaderIP() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    BasicBlock *Preheader = getPreheader();
 
    return {Preheader, std::prev(Preheader->end())};
 
  };
 
 
 
  /// Return the insertion point for user code in the body.
 
  OpenMPIRBuilder::InsertPointTy getBodyIP() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    BasicBlock *Body = getBody();
 
    return {Body, Body->begin()};
 
  };
 
 
 
  /// Return the insertion point for user code after the loop.
 
  OpenMPIRBuilder::InsertPointTy getAfterIP() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    BasicBlock *After = getAfter();
 
    return {After, After->begin()};
 
  };
 
 
 
  Function *getFunction() const {
 
    assert(isValid() && "Requires a valid canonical loop");
 
    return Header->getParent();
 
  }
 
 
 
  /// Consistency self-check.
 
  void assertOK() const;
 
 
 
  /// Invalidate this loop. That is, the underlying IR does not fulfill the
 
  /// requirements of an OpenMP canonical loop anymore.
 
  void invalidate();
 
};
 
 
 
} // end namespace llvm
 
 
 
#endif // LLVM_FRONTEND_OPENMP_OMPIRBUILDER_H